Built-In Zoo Models#

This page lists all of the natively available models in the FiftyOne Model Zoo.

Check out the API reference for complete instructions for using the Model Zoo.


alexnet-imagenet-torch

Classic neural network that recognizes images and helped launch the deep learning revolution

Classification,Embeddings,Logits,Imagenet,PyTorch,Alexnet,Official

centernet-hg104-1024-coco-tf2

CenterNet model from "Objects as Points" with the Hourglass-104 backbone trained on COCO resized to 1024x1024

Detection,Coco,TensorFlow-2,Centernet

centernet-hg104-512-coco-tf2

CenterNet model from "Objects as Points" with the Hourglass-104 backbone trained on COCO resized to 512x512

Detection,Coco,TensorFlow-2,Centernet

centernet-mobilenet-v2-fpn-512-coco-tf2

CenterNet model from "Objects as Points" with the MobileNetV2 backbone trained on COCO resized to 512x512

Detection,Coco,TensorFlow-2,Centernet,Mobilenet

centernet-resnet101-v1-fpn-512-coco-tf2

CenterNet model from "Objects as Points" with the ResNet-101v1 backbone + FPN trained on COCO resized to 512x512

Detection,Coco,TensorFlow-2,Centernet,Resnet

centernet-resnet50-v1-fpn-512-coco-tf2

CenterNet model from "Objects as Points" with the ResNet-50-v1 backbone + FPN trained on COCO resized to 512x512

Detection,Coco,TensorFlow-2,Centernet,Resnet

centernet-resnet50-v2-512-coco-tf2

CenterNet model from "Objects as Points" with the ResNet-50v2 backbone trained on COCO resized to 512x512

Detection,Coco,TensorFlow-2,Centernet,Resnet

classification-transformer-torch

Vision transformer for image classification and custom fine-tuning on specialized datasets

Classification,Logits,Embeddings,PyTorch,Transformers,Official

clip-vit-base32-torch

Understands both images and text together, enabling search and classification using natural language descriptions

Classification,Logits,Embeddings,PyTorch,Clip,Zero-shot,Transformer,Official

convnext-base-224-torch

Base modern CNN with transformer elements for robust visual understanding

Classification,Imagenet,PyTorch,Transformers,Convnext,Official

convnext-large-224-torch

Large modern CNN demonstrating competitive performance with vision transformers

Classification,Imagenet,PyTorch,Transformers,Convnext

convnext-small-224-torch

Small modernized CNN delivering strong accuracy through architectural innovations

Classification,Imagenet,PyTorch,Transformers,Convnext,Official

convnext-tiny-224-torch

Tiny modern CNN bridging traditional convolutions with transformer-inspired improvements

Classification,Imagenet,PyTorch,Transformers,Convnext

convnext-xlarge-224-torch

Extra-large modern CNN maximizing architectural improvements for top accuracy

Classification,Imagenet,PyTorch,Transformers,Convnext,Official

deeplabv3-cityscapes-tf

DeepLabv3+ semantic segmentation model from "Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation" with Xception backbone trained on the Cityscapes dataset

Segmentation,Cityscapes,TensorFlow,Deeplabv3,Legacy

deeplabv3-mnv2-cityscapes-tf

DeepLabv3+ semantic segmentation model from "Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation" with MobileNetV2 backbone trained on the Cityscapes dataset

Segmentation,Cityscapes,TensorFlow,Deeplabv3,Legacy

deeplabv3-resnet101-coco-torch

Labels everyday objects in images pixel by pixel for general scene understanding and analysis

Segmentation,Coco,PyTorch,Resnet,Deeplabv3,Official

deeplabv3-resnet50-coco-torch

Faster version that quickly identifies and labels objects in images for real-time applications

Segmentation,Coco,PyTorch,Resnet,Deeplabv3,Official

densenet121-imagenet-torch

Compact yet powerful classifier that delivers strong results while using minimal computational resources

Classification,Embeddings,Logits,Imagenet,PyTorch,Densenet,Official

densenet161-imagenet-torch

Dense network that achieves high accuracy for image classification and adapts well to new tasks

Classification,Embeddings,Logits,Imagenet,PyTorch,Densenet

densenet169-imagenet-torch

Deeper variant offering improved accuracy while remaining efficient enough for practical deployment

Classification,Embeddings,Logits,Imagenet,PyTorch,Densenet,Official

densenet201-imagenet-torch

Extra-deep model providing the most detailed features for complex image understanding tasks

Classification,Embeddings,Logits,Imagenet,PyTorch,Densenet,Official

depth-estimation-transformer-torch

Hugging Face Transformers model for monocular depth estimation

Depth,PyTorch,Transformers

detection-transformer-torch

Modern object detector that finds items in images without needing complex post-processing steps

Detection,Logits,Embeddings,PyTorch,Transformers,Official

dfine-large-coco-torch

D-FINE Large from "D-FINE: Redefine Regression Task in DETRs as Fine-grained Distribution Refinement" trained on COCO. Achieves 54.0% AP at 124 FPS on T4 GPU.

Detection,Coco,PyTorch,Transformers,Detr,Official

dfine-medium-coco-torch

D-FINE Medium from "D-FINE: Redefine Regression Task in DETRs as Fine-grained Distribution Refinement" trained on COCO. Mid-size real-time object detector.

Detection,Coco,PyTorch,Transformers,Detr,Official

dfine-nano-coco-torch

D-FINE Nano from "D-FINE: Redefine Regression Task in DETRs as Fine-grained Distribution Refinement" trained on COCO. Ultra-lightweight real-time object detector.

Detection,Coco,PyTorch,Transformers,Detr,Official

dfine-small-coco-torch

D-FINE Small from "D-FINE: Redefine Regression Task in DETRs as Fine-grained Distribution Refinement" trained on COCO. Balanced real-time object detector.

Detection,Coco,PyTorch,Transformers,Detr,Official

dfine-xlarge-coco-torch

D-FINE XLarge from "D-FINE: Redefine Regression Task in DETRs as Fine-grained Distribution Refinement" trained on COCO. Achieves 55.8% AP at 78 FPS on T4 GPU.

Detection,Coco,PyTorch,Transformers,Detr,Official

dinov2-vitb14-reg-torch

Enhanced image search model that resists noise and errors for more reliable similarity matching

Embeddings,PyTorch,Dinov2,Transformer,Official

dinov2-vitb14-torch

Creates searchable image fingerprints for finding similar pictures and organizing large photo collections

Embeddings,PyTorch,Dinov2,Transformer,Official

dinov2-vitg14-reg-torch

Highest-capacity dinov2 search model with maximum stability for finding images across massive diverse datasets

Embeddings,PyTorch,Dinov2,Transformer,Official

dinov2-vitg14-torch

Powerful image search engine that handles enormous photo collections with rich detail extraction

Embeddings,PyTorch,Dinov2,Transformer,Official

dinov2-vitl14-reg-torch

Large stable model for finding and grouping similar images across big databases reliably

Embeddings,PyTorch,Dinov2,Transformer,Official

dinov2-vitl14-torch

Large model that creates detailed image fingerprints for advanced search and automatic grouping

Embeddings,PyTorch,Dinov2,Transformer,Official

dinov2-vits14-reg-torch

Compact stable model for image search that runs efficiently on phones and edge devices

Embeddings,PyTorch,Dinov2,Transformer,Official

dinov2-vits14-torch

Small model enabling image search and similarity matching directly on mobile devices

Embeddings,PyTorch,Dinov2,Transformer,Official

efficientdet-d0-512-coco-tf2

EfficientDet-D0 model from "EfficientDet: Scalable and Efficient Object Detection" trained on COCO resized to 512x512

Detection,Coco,TensorFlow-2,Efficientdet

efficientdet-d0-coco-tf1

EfficientDet-D0 model from "EfficientDet: Scalable and Efficient Object Detection" trained on COCO

Detection,Coco,TensorFlow-1,Efficientdet,Legacy

efficientdet-d1-640-coco-tf2

EfficientDet-D1 model from "EfficientDet: Scalable and Efficient Object Detection" trained on COCO resized to 640x640

Detection,Coco,TensorFlow-2,Efficientdet

efficientdet-d1-coco-tf1

EfficientDet-D1 model from "EfficientDet: Scalable and Efficient Object Detection" trained on COCO

Detection,Coco,TensorFlow-1,Efficientdet,Legacy

efficientdet-d2-768-coco-tf2

EfficientDet-D2 model from "EfficientDet: Scalable and Efficient Object Detection" trained on COCO resized to 768x768

Detection,Coco,TensorFlow-2,Efficientdet

efficientdet-d2-coco-tf1

EfficientDet-D2 model from "EfficientDet: Scalable and Efficient Object Detection" trained on COCO

Detection,Coco,TensorFlow-1,Efficientdet,Legacy

efficientdet-d3-896-coco-tf2

EfficientDet-D3 model from "EfficientDet: Scalable and Efficient Object Detection" trained on COCO resized to 896x896

Detection,Coco,TensorFlow-2,Efficientdet

efficientdet-d3-coco-tf1

EfficientDet-D3 model from "EfficientDet: Scalable and Efficient Object Detection" trained on COCO

Detection,Coco,TensorFlow-1,Efficientdet,Legacy

efficientdet-d4-1024-coco-tf2

EfficientDet-D4 model from "EfficientDet: Scalable and Efficient Object Detection" trained on COCO resized to 1024x1024

Detection,Coco,TensorFlow-2,Efficientdet

efficientdet-d4-coco-tf1

EfficientDet-D4 model from "EfficientDet: Scalable and Efficient Object Detection" trained on COCO

Detection,Coco,TensorFlow-1,Efficientdet,Legacy

efficientdet-d5-1280-coco-tf2

EfficientDet-D5 model from "EfficientDet: Scalable and Efficient Object Detection" trained on COCO resized to 1280x1280

Detection,Coco,TensorFlow-2,Efficientdet

efficientdet-d5-coco-tf1

EfficientDet-D5 model from "EfficientDet: Scalable and Efficient Object Detection" trained on COCO

Detection,Coco,TensorFlow-1,Efficientdet,Legacy

efficientdet-d6-1280-coco-tf2

EfficientDet-D6 model from "EfficientDet: Scalable and Efficient Object Detection" trained on COCO resized to 1280x1280

Detection,Coco,TensorFlow-2,Efficientdet

efficientdet-d6-coco-tf1

EfficientDet-D6 model from "EfficientDet: Scalable and Efficient Object Detection" trained on COCO

Detection,Coco,TensorFlow-1,Efficientdet,Legacy

efficientdet-d7-1536-coco-tf2

EfficientDet-D7 model from "EfficientDet: Scalable and Efficient Object Detection" trained on COCO resized to 1536x1536

Detection,Coco,TensorFlow-2,Efficientdet

efficientnet-b0-imagenet-torch

Efficient image classifier optimized for mobile devices with excellent accuracy-efficiency tradeoff

Classification,Imagenet,PyTorch,Transformers,Efficientnet,Official

efficientnet-b1-imagenet-torch

Scaled efficient classifier with improved accuracy for slightly larger computational budgets

Classification,Imagenet,PyTorch,Transformers,Efficientnet,Official

efficientnet-b2-imagenet-torch

Balanced efficient model providing stronger performance while maintaining reasonable resource usage

Classification,Imagenet,PyTorch,Transformers,Efficientnet,Official

efficientnet-b3-imagenet-torch

Mid-scale efficient classifier delivering high accuracy for versatile deployment scenarios

Classification,Imagenet,PyTorch,Transformers,Efficientnet,Official

efficientnet-b4-imagenet-torch

Large efficient model with enhanced features for transfer learning applications

Classification,Imagenet,PyTorch,Transformers,Efficientnet,Official

efficientnet-b5-imagenet-torch

High-capacity efficient classifier prioritizing accuracy with available compute resources

Classification,Imagenet,PyTorch,Transformers,Efficientnet,Official

efficientnet-b6-imagenet-torch

Extended efficient model approaching state-of-the-art accuracy on challenging datasets

Classification,Imagenet,PyTorch,Transformers,Efficientnet,Official

efficientnet-b7-imagenet-torch

Maximum efficient classifier pushing performance boundaries while preserving efficiency principles

Classification,Imagenet,PyTorch,Transformers,Efficientnet,Official

faster-rcnn-inception-resnet-atrous-v2-coco-tf

Faster R-CNN model from "Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks" atrous version with Inception backbone trained on COCO

Detection,Coco,TensorFlow,Faster-rcnn,Inception,Resnet

faster-rcnn-inception-resnet-atrous-v2-lowproposals-coco-tf

Faster R-CNN model from "Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks" atrous version with low-proposals and Inception backbone trained on COCO

Detection,Coco,TensorFlow,Faster-rcnn,Inception,Resnet,Legacy

faster-rcnn-inception-v2-coco-tf

Faster R-CNN model from "Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks" with Inception v2 backbone trained on COCO

Detection,Coco,TensorFlow,Faster-rcnn,Inception

faster-rcnn-nas-coco-tf

Faster R-CNN model from "Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks" with NAS-net backbone trained on COCO

Detection,Coco,TensorFlow,Faster-rcnn

faster-rcnn-nas-lowproposals-coco-tf

Faster R-CNN model from "Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks" with low-proposals and NAS-net backbone trained on COCO

Detection,Coco,TensorFlow,Faster-rcnn,Legacy

faster-rcnn-resnet101-coco-tf

Faster R-CNN model from "Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks" with ResNet-101 backbone trained on COCO

Detection,Coco,TensorFlow,Faster-rcnn,Resnet,Legacy

faster-rcnn-resnet101-lowproposals-coco-tf

Faster R-CNN model from "Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks" with low-proposals and ResNet-101 backbone trained on COCO

Detection,Coco,TensorFlow,Faster-rcnn,Resnet,Legacy

faster-rcnn-resnet50-coco-tf

Faster R-CNN model from "Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks" with ResNet-50 backbone trained on COCO

Detection,Coco,TensorFlow,Faster-rcnn,Resnet

faster-rcnn-resnet50-fpn-coco-torch

Multi-scale object finder that accurately detects both small and large items in images

Detection,Coco,PyTorch,Faster-rcnn,Resnet,Official

faster-rcnn-resnet50-lowproposals-coco-tf

Faster R-CNN model from "Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks" with low-proposals and ResNet-50 backbone trained on COCO

Detection,Coco,TensorFlow,Faster-rcnn,Resnet,Legacy

fcn-resnet101-coco-torch

Creates detailed pixel-level labels for images, identifying and outlining twenty-one different object categories

Segmentation,Coco,PyTorch,Fcn,Resnet,Official

fcn-resnet50-coco-torch

Fast image labeler that quickly identifies and outlines objects for interactive editing and annotation

Segmentation,Coco,PyTorch,Fcn,Resnet,Official

googlenet-imagenet-torch

Classic image classifier providing reliable categorization and features for various computer vision projects.

Classification,Embeddings,Logits,Imagenet,PyTorch,Googlenet,Official

group-vit-segmentation-transformer-torch

Hugging Face Transformers model for zero-shot semantic segmentation

Segmentation,Embeddings,PyTorch,Transformers,Zero-shot,Official

inception-resnet-v2-imagenet-tf1

Inception v2 model from "Rethinking the Inception Architecture for Computer Vision" trained on ImageNet

Classification,Embeddings,Logits,Imagenet,TensorFlow-1,Inception,Resnet

inception-v3-imagenet-torch

Efficient image classifier delivering accurate results with useful features for transfer learning applications

Classification,Embeddings,Logits,Imagenet,PyTorch,Inception,Official

inception-v4-imagenet-tf1

Inception v4 model from "Inception-v4, Inception-ResNet and the Impact of Residual Connections on Learning" trained on ImageNet

Classification,Embeddings,Logits,Imagenet,TensorFlow-1,Inception,Legacy

keypoint-rcnn-resnet50-fpn-coco-torch

Finds people in images and maps their body joints for pose estimation and motion analysis

Keypoints,Coco,PyTorch,Keypoint-rcnn,Resnet,Official

mask-rcnn-inception-resnet-v2-atrous-coco-tf

Mask R-CNN model from "Mask R-CNN" atrous version with Inception backbone trained on COCO

Instances,Coco,TensorFlow,Mask-rcnn,Inception,Resnet

mask-rcnn-inception-v2-coco-tf

Mask R-CNN model from "Mask R-CNN" with Inception backbone trained on COCO

Instances,Coco,TensorFlow,Mask-rcnn,Inception

mask-rcnn-resnet101-atrous-coco-tf

Mask R-CNN model from "Mask R-CNN" atrous version with ResNet-101 backbone trained on COCO

Instances,Coco,TensorFlow,Mask-rcnn,Resnet,Legacy

mask-rcnn-resnet50-atrous-coco-tf

Mask R-CNN model from "Mask R-CNN" atrous version with ResNet-50 backbone trained on COCO

Instances,Coco,TensorFlow,Mask-rcnn,Resnet,Legacy

mask-rcnn-resnet50-fpn-coco-torch

Multi-scale object outliner using advanced architecture for accurate segmentation across different object sizes

Instances,Coco,PyTorch,Mask-rcnn,Resnet,Official

med-sam-2-video-torch

Medical segmentation tool that outlines organs and structures in medical videos and 3D scans

Segment-anything,PyTorch,Zero-shot,Video,Med-sam,Transformer,Official

mnasnet0.5-imagenet-torch

Ultra-lightweight image classifier designed by AI for running directly on phones and IoT devices

Classification,Embeddings,Logits,Imagenet,PyTorch,Mnasnet,Official

mnasnet1.0-imagenet-torch

Mobile-optimized classifier balancing size and accuracy for efficient on-device image recognition

Classification,Embeddings,Logits,Imagenet,PyTorch,Mnasnet,Official

mobilenet-v2-imagenet-tf1

MobileNetV2 model from "MobileNetV2: Inverted Residuals and Linear Bottlenecks" trained on ImageNet

Classification,Embeddings,Logits,Imagenet,TensorFlow-1,Mobilenet

mobilenet-v2-imagenet-torch

Mobile-friendly image classifier optimized for quick training and deployment on resource-limited devices

Classification,Embeddings,Logits,Imagenet,PyTorch,Mobilenet,Official

omdet-turbo-swin-tiny-torch

Real-time detector that finds any object you describe in words, perfect for live video analysis

Detection,Logits,Embeddings,PyTorch,Transformers,Zero-shot,Official

open-clip-torch

Connects images with text descriptions enabling search by words and automatic content filtering systems

Classification,Logits,Embeddings,PyTorch,Clip,Zero-shot,Transformer

openbmb/MiniCPM-V-4_5

MiniCPM-V 4.5 is the latest and most capable model in the MiniCPM-V series. The model is built on Qwen3-8B and SigLIP2-400M with a total of 8B parameters.

Detection,Keypoints,Ocr,Vlm,Classification,Zero-shot

owlvit-base-patch16-torch

Finds any object you name in pictures using 16x16 image patches without needing specific training for those items

Detection,Logits,Embeddings,PyTorch,Transformers,Zero-shot,Official

owlvit-base-patch32-torch

Finds any object you name in pictures using efficient 32x32 image patches without needing specific training

Detection,Logits,Embeddings,PyTorch,Transformers,Zero-shot,Official

owlvit-large-patch14-torch

Large OWL-ViT zero-shot object detector with ViT-L/14 backbone. Achieves higher accuracy than base models, especially for smaller objects.

Detection,Logits,Embeddings,PyTorch,Transformers,Zero-shot,Official

resnet-v1-50-imagenet-tf1

ResNet-50 v1 model from "Deep Residual Learning for Image Recognition" trained on ImageNet

Classification,Embeddings,Logits,Imagenet,TensorFlow-1,Resnet,Legacy

resnet-v2-50-imagenet-tf1

ResNet-50 v2 model from "Deep Residual Learning for Image Recognition" trained on ImageNet

Classification,Embeddings,Logits,Imagenet,TensorFlow-1,Resnet,Legacy

resnet101-imagenet-torch

Deep image recognition model delivering high accuracy for demanding classification and analysis tasks

Classification,Embeddings,Logits,Imagenet,PyTorch,Resnet,Official

resnet152-imagenet-torch

Very deep classifier providing the richest visual features for precision-critical image understanding applications

Classification,Embeddings,Logits,Imagenet,PyTorch,Resnet,Official

resnet18-imagenet-torch

Lightweight image classifier designed for fast recognition on phones and other resource-limited devices

Classification,Embeddings,Logits,Imagenet,PyTorch,Resnet,Official

resnet34-imagenet-torch

Balanced image classifier offering good accuracy and speed for everyday computer vision needs

Classification,Embeddings,Logits,Imagenet,PyTorch,Resnet,Official

resnet50-imagenet-torch

Most popular image recognition backbone widely used as starting point for custom vision projects

Classification,Embeddings,Logits,Imagenet,PyTorch,Resnet,Official

resnext101-32x8d-imagenet-torch

Powerful image classifier with enhanced capacity for handling complex visual recognition challenges effectively

Classification,Embeddings,Logits,Imagenet,PyTorch,Resnext,Official

resnext50-32x4d-imagenet-torch

Efficient advanced classifier delivering strong accuracy with reasonable computing requirements for practical deployments

Classification,Embeddings,Logits,Imagenet,PyTorch,Resnext

retinanet-resnet50-fpn-coco-torch

Fast object detector that quickly finds and boxes eighty common items in any image

Detection,Coco,PyTorch,Retinanet,Resnet

rfcn-resnet101-coco-tf

R-FCN object detection model from "R-FCN: Object Detection via Region-based Fully Convolutional Networks" with ResNet-101 backbone trained on COCO

Detection,Coco,TensorFlow,Rfcn,Resnet,Legacy

rtdetr-l-coco-torch

Modern real-time object detector that finds items without complex post-processing for responsive applications

Detection,Coco,PyTorch,Transformer,Rtdetr,Official

rtdetr-v2-m-coco-torch

Balanced real-time object detector offering improved accuracy for production use

Detection,Coco,PyTorch,Transformers,Rtdetr,Official

rtdetr-v2-s-coco-torch

Lightweight real-time object detector optimized for speed on edge devices

Detection,Coco,PyTorch,Transformers,Rtdetr

rtdetr-x-coco-torch

High-capacity object detector delivering very precise results at speeds suitable for production use

Detection,Coco,PyTorch,Transformer,Rtdetr,Official

segformer-b0-ade20k-torch

Efficient transformer-based semantic segmentation model for scene parsing with 150 classes

Segmentation,PyTorch,Segformer,Official

segformer-b1-ade20k-torch

Balanced SegFormer model providing good accuracy-efficiency tradeoff for scene understanding

Segmentation,PyTorch,Segformer,Official

segformer-b2-ade20k-torch

Medium-sized SegFormer delivering enhanced segmentation quality for complex scenes

Segmentation,PyTorch,Segformer,Official

segformer-b3-ade20k-torch

Larger SegFormer model with improved accuracy for detailed semantic segmentation

Segmentation,PyTorch,Segformer,Official

segformer-b4-ade20k-torch

High-capacity SegFormer achieving excellent results on challenging segmentation tasks

Segmentation,PyTorch,Segformer,Official

segformer-b5-ade20k-torch

Largest SegFormer model delivering the best semantic segmentation performance in its family

Segmentation,PyTorch,Segformer,Official

segment-anything-2-hiera-base-plus-image-torch

Accurate image segmentation model for editing, labeling, and creative work with still pictures

Segment-anything,PyTorch,Zero-shot,Transformer,Official

segment-anything-2-hiera-base-plus-video-torch

Video segmentation model that tracks and outlines objects throughout clips for editing and analysis

Segment-anything,PyTorch,Zero-shot,Video,Transformer,Official

segment-anything-2-hiera-large-image-torch

High-quality image segmenter producing detailed masks for demanding professional editing and annotation tasks

Segment-anything,PyTorch,Zero-shot,Transformer,Official

segment-anything-2-hiera-large-video-torch

Advanced video segmenter providing fine object tracking throughout full videos for post-production work

Segment-anything,PyTorch,Zero-shot,Video,Transformer

segment-anything-2-hiera-small-image-torch

Fast image segmentation model that runs efficiently on laptops and edge computing devices

Segment-anything,PyTorch,Zero-shot,Transformer,Official

segment-anything-2-hiera-small-video-torch

Quick video segmentation model delivering rapid object tracking on standard graphics cards

Segment-anything,PyTorch,Zero-shot,Video,Transformer,Official

segment-anything-2-hiera-tiny-image-torch

Smallest image segmentation model offering instant results for mobile apps and embedded systems

Segment-anything,PyTorch,Zero-shot

segment-anything-2-hiera-tiny-video-torch

Tiny video segmenter enabling real-time object tracking on phones and compact devices

Segment-anything,PyTorch,Zero-shot,Video,Transformer,Official

segment-anything-2.1-hiera-base-plus-image-torch

Updated image segmenter with improved mask accuracy for everyday editing and dataset creation

Segment-anything,PyTorch,Zero-shot,Transformer,Official

segment-anything-2.1-hiera-base-plus-video-torch

Enhanced video segmenter with better tracking quality for video analysis and scene understanding

Segment-anything,PyTorch,Zero-shot,Video,Transformer,Official

segment-anything-2.1-hiera-large-image-torch

Large updated model offering even finer masks for high-resolution professional image workflows

Segment-anything,PyTorch,Zero-shot,Transformer,Official

segment-anything-2.1-hiera-large-video-torch

Large video model producing exceptionally detailed masks throughout long videos for intensive production

Segment-anything,PyTorch,Zero-shot,Video,Transformer,Official

segment-anything-2.1-hiera-small-image-torch

Balanced updated segmenter combining speed and accuracy for edge device image processing

Segment-anything,PyTorch,Zero-shot,Transformer,Official

segment-anything-2.1-hiera-small-video-torch

Improved video segmenter maintaining quick performance on compact hardware while enhancing mask quality

Segment-anything,PyTorch,Zero-shot,Video,Transformer,Official

segment-anything-2.1-hiera-tiny-image-torch

Enhanced mobile image segmenter for apps, augmented reality filters, and on-device processing

Segment-anything,PyTorch,Zero-shot,Transformer

segment-anything-2.1-hiera-tiny-video-torch

Upgraded mobile video segmenter for live effects on phones, wearables, and smart cameras

Segment-anything,PyTorch,Zero-shot,Video,Transformer

segment-anything-vitb-torch

Interactive segmentation tool that instantly outlines any object you point to or describe

Segment-anything,Sa-1b,PyTorch,Zero-shot,Transformer,Official

segment-anything-vith-torch

Highest quality segmentation model creating extremely detailed masks for research and large-scale annotation projects

Segment-anything,Sa-1b,PyTorch,Zero-shot,Transformer,Official

segment-anything-vitl-torch

Large segmentation model producing finer object outlines for professional editing and labeling workflows

Segment-anything,Sa-1b,PyTorch,Zero-shot,Transformer

segmentation-transformer-torch

Hugging Face Transformers model for semantic segmentation

Segmentation,PyTorch,Transformers,Official

shufflenetv2-0.5x-imagenet-torch

Ultra-small image classifier for tiny devices with very limited power and memory

Classification,Embeddings,Logits,Imagenet,PyTorch,Shufflenet,Official

shufflenetv2-1.0x-imagenet-torch

Mobile image classifier that works efficiently on phones with modest computing resources

Classification,Embeddings,Logits,Imagenet,PyTorch,Shufflenet,Official

siglip-base-patch16-224-torch

Hugging Face Transformers model for zero-shot image classification

Classification,Logits,Embeddings,PyTorch,Transformers,Zero-shot,Official

squeezenet-1.1-imagenet-torch

Tiny image classifier that fits in just five megabytes for embedded devices

Classification,Imagenet,PyTorch,Squeezenet

squeezenet-imagenet-torch

Ultra-compact image classifier perfect for severely resource-constrained hardware and applications

Classification,Imagenet,PyTorch,Squeezenet,Official

ssd-inception-v2-coco-tf

Inception Single Shot Detector model from "SSD: Single Shot MultiBox Detector" trained on COCO

Detection,Coco,TensorFlow,Ssd,Inception

ssd-mobilenet-v1-coco-tf

Single Shot Detector model from "SSD: Single Shot MultiBox Detector" with MobileNetV1 backbone trained on COCO

Detection,Coco,TensorFlow,Ssd,Mobilenet,Legacy

ssd-mobilenet-v1-fpn-640-coco17

MobileNetV1 model from "MobileNetV2: Inverted Residuals and Linear Bottlenecks" resized to 640x640

Detection,Coco,TensorFlow-2,Ssd,Mobilenet

ssd-mobilenet-v1-fpn-coco-tf

FPN Single Shot Detector model from "SSD: Single Shot MultiBox Detector" with MobileNetV1 backbone trained on COCO

Detection,Coco,TensorFlow,Ssd,Mobilenet,Legacy

ssd-mobilenet-v2-320-coco17

MobileNetV2 model from "MobileNetV2: Inverted Residuals and Linear Bottlenecks" resized to 320x320

Detection,Coco,TensorFlow-2,Ssd,Mobilenet

ssd-resnet50-fpn-coco-tf

FPN Single Shot Detector model from "SSD: Single Shot MultiBox Detector" with ResNet-50 backbone trained on COCO

Detection,Coco,TensorFlow,Ssd,Resnet,Legacy

swin-v2-base-torch

Base hierarchical transformer delivering strong results across vision tasks

Classification,Imagenet,PyTorch,Transformers,Swin-transformer,Official

swin-v2-large-torch

Large hierarchical transformer with enhanced capacity for demanding applications

Classification,Imagenet,PyTorch,Transformers,Swin-transformer,Official

swin-v2-small-torch

Small hierarchical transformer balancing efficiency and performance for practical use

Classification,Imagenet,PyTorch,Transformers,Swin-transformer,Official

swin-v2-tiny-torch

Tiny hierarchical transformer for efficient visual recognition on edge devices

Classification,Imagenet,PyTorch,Transformers,Swin-transformer,Official

vgg11-bn-imagenet-torch

Classic image classifier with stable training useful for various computer vision projects

Classification,Embeddings,Logits,Imagenet,PyTorch,Vgg,Official

vgg11-imagenet-torch

Simple baseline image classifier valuable for research experimentation and learning purposes

Classification,Embeddings,Logits,Imagenet,PyTorch,Vgg,Official

vgg13-bn-imagenet-torch

Deeper classic classifier providing stable training process and solid accuracy results overall

Classification,Embeddings,Logits,Imagenet,PyTorch,Vgg,Official

vgg13-imagenet-torch

Straightforward image classifier valued for easy experimentation and model compression studies

Classification,Embeddings,Logits,Imagenet,PyTorch,Vgg,Official

vgg16-bn-imagenet-torch

Popular feature extractor widely used for detection, style transfer, and medical imaging

Classification,Embeddings,Logits,Imagenet,PyTorch,Vgg,Official

vgg16-imagenet-tf1

VGG-16 model from "Very Deep Convolutional Networks for Large-Scale Image Recognition" trained on ImageNet

Classification,Embeddings,Logits,Imagenet,TensorFlow-1,Vgg,Legacy

vgg16-imagenet-torch

PyTorch version of the popular classifier ready for modern deep learning workflows

Classification,Embeddings,Logits,Imagenet,PyTorch,Vgg,Official

vgg19-bn-imagenet-torch

Deep classic model providing rich features for style transfer and interpretability analysis

Classification,Embeddings,Logits,Imagenet,PyTorch,Vgg,Official

vgg19-imagenet-torch

Deep image classifier delivering detailed features for creative applications and research projects

Classification,Embeddings,Logits,Imagenet,PyTorch,Vgg,Official

vit-base-patch16-224-imagenet-torch

Modern image classifier that recognizes objects and provides useful features for various computer vision tasks

Classification,Logits,Embeddings,PyTorch,Transformers,Official

wide-resnet101-2-imagenet-torch

Extra-wide deep classifier for high-precision image recognition and advanced transfer learning tasks

Classification,Embeddings,Logits,Imagenet,PyTorch,Wide-resnet,Official

wide-resnet50-2-imagenet-torch

Wide classifier offering stronger accuracy and better features for adapting to new tasks

Classification,Embeddings,Logits,Imagenet,PyTorch,Wide-resnet,Official

yolo-nas-torch

AI-designed detector family offering three model variants for diverse deployment scenarios

Detection,PyTorch,Yolo,Official

yolo-v2-coco-tf1

YOLOv2 model from "YOLO9000: Better, Faster, Stronger" trained on COCO

Detection,Coco,TensorFlow-1,Yolo,Legacy

yolo11l-coco-torch

Real-time object detector balancing high accuracy with fast processing speeds effectively

Detection,Coco,PyTorch,Yolo,Official

yolo11l-seg-coco-torch

Model creating detailed object outlines for precise image editing and analysis

Instances,Coco,PyTorch,Yolo,Official

yolo11m-coco-torch

Object detector offering good balance between speed and accuracy for most applications

Detection,Coco,PyTorch,Yolo,Official

yolo11m-seg-coco-torch

Model generating object masks efficiently for everyday segmentation tasks

Instances,Coco,PyTorch,Yolo,Official

yolo11n-coco-torch

Object detector designed specifically for phones and other edge computing devices

Detection,Coco,PyTorch,Yolo,Official

yolo11n-seg-coco-torch

Edge model producing object outlines directly on phones and edge devices

Instances,Coco,PyTorch,Yolo,Official

yolo11s-coco-torch

Fast object detector ideal for systems with limited graphics processing power

Detection,Coco,PyTorch,Yolo,Official

yolo11s-seg-coco-torch

Model creating object masks quickly for real-time segmentation applications

Instances,Coco,PyTorch,Yolo,Official

yolo11x-coco-torch

Object detector prioritizing accuracy over processing speed for critical applications

Detection,Coco,PyTorch,Yolo,Official

yolo11x-seg-coco-torch

Model delivering high-quality object outlines for professional workflows

Instances,Coco,PyTorch,Yolo,Official

yoloe11l-seg-torch

Real-time model creating both object outlines and boxes for any described item

Instances,PyTorch,Yolo,Zero-shot,Official

yoloe11m-seg-torch

Model producing masks and boxes for objects described in natural language

Instances,PyTorch,Yolo,Zero-shot,Official

yoloe11s-seg-torch

Segments specified classes, generating object outlines and boxes for real-time applications

Instances,PyTorch,Yolo,Zero-shot,Official

yoloev8l-seg-torch

Model outlining and boxing any object you describe without specific training

Instances,PyTorch,Yolo,Zero-shot,Official

yoloev8m-seg-torch

Model creating masks for any object type you name in text

Instances,PyTorch,Yolo,Zero-shot,Official

yoloev8s-seg-torch

Compact model producing outlines for objects described in words on edge devices

Instances,PyTorch,Yolo,Zero-shot,Official

yolov10l-coco-torch

Object detector with special optimizations for even faster inference on modern hardware

Detection,Coco,PyTorch,Yolo,Official

yolov10m-coco-torch

Balanced detector providing good accuracy and speed for general-purpose object detection tasks

Detection,Coco,PyTorch,Yolo,Official

yolov10n-coco-torch

Edge-optimized detector for devices with minimal computing resources available

Detection,Coco,PyTorch,Yolo,Official

yolov10s-coco-torch

Fast lightweight detector suitable for systems with limited GPU capabilities and memory

Detection,Coco,PyTorch,Yolo,Official

yolov10x-coco-torch

High-accuracy detector for demanding object detection applications and research

Detection,Coco,PyTorch,Yolo,Official

yolov5l-coco-torch

Real-time detector producing accurate results quickly for demanding vision applications

Detection,Coco,PyTorch,Yolo,Official

yolov5m-coco-torch

Real-time detector balancing good accuracy with fast processing speeds

Detection,Coco,PyTorch,Yolo,Official

yolov5n-coco-torch

Lightweight detector for edge devices needing basic object detection capabilities

Detection,Coco,PyTorch,Yolo,Official

yolov5s-coco-torch

Real-time detector delivering good results with minimal computational requirements

Detection,Coco,PyTorch,Yolo,Official

yolov5x-coco-torch

High-accuracy detector offering top precision for applications where quality is critical

Detection,Coco,PyTorch,Yolo,Official

yolov8l-coco-torch

Real-time detector with advanced architecture for improved object finding in complex scenes

Detection,Coco,PyTorch,Yolo,Official

yolov8l-obb-dotav1-torch

Specialized detector that finds rotated objects in aerial and satellite imagery accurately

Detection,PyTorch,Yolo,Polylines,Obb,Official

yolov8l-oiv7-torch

General-purpose detector trained on diverse images recognizing over six hundred object categories

Detection,Oiv7,PyTorch,Yolo,Official

yolov8l-seg-coco-torch

Creates precise object outlines for detailed image editing and analysis tasks

Instances,Coco,PyTorch,Yolo,Official

yolov8l-world-torch

Finds and boxes any object you describe using natural language prompts

Detection,PyTorch,Yolo,Zero-shot,Official

yolov8m-coco-torch

Detector balancing speed and accuracy for everyday object detection needs

Detection,Coco,PyTorch,Yolo,Official

yolov8m-obb-dotav1-torch

Finds rotated bounding boxes in aerial images for mapping and surveillance applications

Detection,PyTorch,Yolo,Polylines,Obb,Official

yolov8m-oiv7-torch

Versatile detector recognizing hundreds of different object types across varied image domains

Detection,Oiv7,PyTorch,Yolo,Official

yolov8m-seg-coco-torch

Generates object masks with good balance of speed and quality

Instances,Coco,PyTorch,Yolo,Official

yolov8m-world-torch

Detector understanding text descriptions to find matching objects in images

Detection,PyTorch,Yolo,Zero-shot,Official

yolov8n-coco-torch

Edge-optimized detector recognizing common objects on resource-limited devices effectively

Detection,Coco,PyTorch,Yolo,Official

yolov8n-obb-dotav1-torch

Lightweight detector for finding rotated objects in aerial imagery on edge hardware

Detection,PyTorch,Yolo,Polylines,Obb,Official

yolov8n-oiv7-torch

Edge-friendly detector recognizing hundreds of object categories on resource-limited devices effectively

Detection,Oiv7,PyTorch,Yolo,Official

yolov8n-seg-coco-torch

Edge-optimized model producing object outlines on devices with limited resources.

Instances,Coco,PyTorch,Yolo,Official

yolov8s-coco-torch

Detector offering fast performance on mid-range graphics cards and processors

Detection,Coco,PyTorch,Yolo,Official

yolov8s-obb-dotav1-torch

Efficiently finds rotated objects in aerial photos for mapping and analysis tasks

Detection,PyTorch,Yolo,Polylines,Obb,Official

yolov8s-oiv7-torch

Compact detector recognizing diverse object types across many different image categories

Detection,Oiv7,PyTorch,Yolo,Official

yolov8s-seg-coco-torch

Fast model creating object masks for real-time image segmentation needs

Instances,Coco,PyTorch,Yolo,Official

yolov8s-world-torch

Lightweight detector finding objects based on text descriptions for edge applications

Detection,PyTorch,Yolo,Zero-shot,Official

yolov8x-coco-torch

High-accuracy detector for critical applications where precision matters most

Detection,Coco,PyTorch,Yolo,Official

yolov8x-obb-dotav1-torch

High-precision detector for rotated objects in aerial and satellite imagery analysis

Detection,PyTorch,Yolo,Polylines,Obb,Official

yolov8x-oiv7-torch

Accurate general detector recognizing over six hundred different object types

Detection,Oiv7,PyTorch,Yolo,Official

yolov8x-seg-coco-torch

High-accuracy model generating detailed object outlines for demanding professional applications

Instances,Coco,PyTorch,Yolo,Official

yolov8x-world-torch

Open-vocabulary detector with high accuracy for text-based object finding

Detection,PyTorch,Yolo,Zero-shot,Official

yolov9c-coco-torch

Detector enhanced with transformer technology for improved object finding capabilities

Detection,Coco,PyTorch,Yolo,Official

yolov9c-seg-coco-torch

Compact model producing both masks and boxes with transformer-enhanced accuracy

Instances,Coco,PyTorch,Yolo,Official

yolov9e-coco-torch

Advanced detector with transformer backbone delivering superior accuracy for complex scenes

Detection,Coco,PyTorch,Yolo,Official

yolov9e-seg-coco-torch

Advanced model creating precise object outlines using enhanced transformer architecture

Instances,Coco,PyTorch,Yolo,Official

zero-shot-classification-transformer-torch

Finds any object you name in images without requiring training on those specific items

Classification,Logits,Embeddings,PyTorch,Transformers,Zero-shot,Official

zero-shot-detection-transformer-torch

Hugging Face Transformers model for zero-shot object detection

Detection,Logits,Embeddings,PyTorch,Transformers,Zero-shot,Official

Torch models#

alexnet-imagenet-torch#

Classic neural network that recognizes images and helped launch the deep learning revolution.

Details

  • Model name: alexnet-imagenet-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Alex Krizhevsky

  • Model license: BSD 3-Clause

  • Model size: 233.10 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, torch, alexnet, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("alexnet-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

classification-transformer-torch#

Vision transformer for image classification and custom fine-tuning on specialized datasets.

Details

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("classification-transformer-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

clip-vit-base32-torch#

Understands both images and text together, enabling search and classification using natural language descriptions.

Details

  • Model name: clip-vit-base32-torch

  • Model source: openai/CLIP

  • Model author: Alec Radford, et al.

  • Model license: MIT

  • Model size: 337.58 MB

  • Exposes embeddings? yes

  • Tags: classification, logits, embeddings, torch, clip, zero-shot, transformer, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("clip-vit-base32-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
17
18#
19# Make zero-shot predictions with custom classes
20#
21
22model = foz.load_zoo_model(
23    "clip-vit-base32-torch",
24    text_prompt="A photo of a",
25    classes=["person", "dog", "cat", "bird", "car", "tree", "chair"],
26)
27
28dataset.apply_model(model, label_field="predictions")
29session.refresh()

convnext-base-224-torch#

Base modern CNN with transformer elements for robust visual understanding.

Details

  • Model name: convnext-base-224-torch

  • Model source: https://huggingface.co/facebook/convnext-base-224

  • Model author: Zhuang Liu, et al.

  • Model license: Apache 2.0

  • Model size: 1.32 GB

  • Exposes embeddings? no

  • Tags: classification, imagenet, torch, transformers, convnext, official

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("convnext-base-224-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

convnext-large-224-torch#

Large modern CNN demonstrating competitive performance with vision transformers.

Details

  • Model name: convnext-large-224-torch

  • Model source: https://huggingface.co/facebook/convnext-large-224

  • Model author: Zhuang Liu, et al.

  • Model license: Apache 2.0

  • Model size: 2.95 GB

  • Exposes embeddings? no

  • Tags: classification, imagenet, torch, transformers, convnext

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("convnext-large-224-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

convnext-small-224-torch#

Small modernized CNN delivering strong accuracy through architectural innovations.

Details

  • Model name: convnext-small-224-torch

  • Model source: https://huggingface.co/facebook/convnext-small-224

  • Model author: Zhuang Liu, et al.

  • Model license: Apache 2.0

  • Model size: 383.39 MB

  • Exposes embeddings? no

  • Tags: classification, imagenet, torch, transformers, convnext, official

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("convnext-small-224-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

convnext-tiny-224-torch#

Tiny modern CNN bridging traditional convolutions with transformer-inspired improvements.

Details

  • Model name: convnext-tiny-224-torch

  • Model source: https://huggingface.co/facebook/convnext-tiny-224

  • Model author: Zhuang Liu, et al.

  • Model license: Apache 2.0

  • Model size: 436.54 MB

  • Exposes embeddings? no

  • Tags: classification, imagenet, torch, transformers, convnext

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("convnext-tiny-224-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

convnext-xlarge-224-torch#

Extra-large modern CNN maximizing architectural improvements for top accuracy.

Details

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("convnext-xlarge-224-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

deeplabv3-resnet101-coco-torch#

Labels everyday objects in images pixel by pixel for general scene understanding and analysis.

Details

  • Model name: deeplabv3-resnet101-coco-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Liang-Chieh Chen, et al.

  • Model license: BSD 3-Clause

  • Model size: 233.22 MB

  • Exposes embeddings? no

  • Tags: segmentation, coco, torch, resnet, deeplabv3, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("deeplabv3-resnet101-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

deeplabv3-resnet50-coco-torch#

Faster version that quickly identifies and labels objects in images for real-time applications.

Details

  • Model name: deeplabv3-resnet50-coco-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Liang-Chieh Chen, et al.

  • Model license: BSD 3-Clause

  • Model size: 160.51 MB

  • Exposes embeddings? no

  • Tags: segmentation, coco, torch, resnet, deeplabv3, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("deeplabv3-resnet50-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

densenet121-imagenet-torch#

Compact yet powerful classifier that delivers strong results while using minimal computational resources.

Details

  • Model name: densenet121-imagenet-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Gao Huang, et al.

  • Model license: BSD 3-Clause

  • Model size: 30.84 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, torch, densenet, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("densenet121-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

densenet161-imagenet-torch#

Dense network that achieves high accuracy for image classification and adapts well to new tasks.

Details

  • Model name: densenet161-imagenet-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Gao Huang, et al.

  • Model license: BSD 3-Clause

  • Model size: 110.37 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, torch, densenet

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("densenet161-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

densenet169-imagenet-torch#

Deeper variant offering improved accuracy while remaining efficient enough for practical deployment.

Details

  • Model name: densenet169-imagenet-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Gao Huang, et al.

  • Model license: BSD 3-Clause

  • Model size: 54.71 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, torch, densenet, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("densenet169-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

densenet201-imagenet-torch#

Extra-deep model providing the most detailed features for complex image understanding tasks.

Details

  • Model name: densenet201-imagenet-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Gao Huang, et al.

  • Model license: BSD 3-Clause

  • Model size: 77.37 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, torch, densenet, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("densenet201-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

depth-estimation-transformer-torch#

Hugging Face Transformers model for monocular depth estimation.

Details

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("depth-estimation-transformer-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

detection-transformer-torch#

Modern object detector that finds items in images without needing complex post-processing steps.

Details

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("detection-transformer-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

dfine-large-coco-torch#

D-FINE Large from D-FINE: Redefine Regression Task in DETRs as Fine-grained Distribution Refinement trained on COCO. Achieves 54.0% AP at 124 FPS on T4 GPU..

Details

  • Model name: dfine-large-coco-torch

  • Model source: Peterande/D-FINE

  • Model author: Yansong Peng, et al.

  • Model license: Apache 2.0

  • Model size: 50.00 MB

  • Exposes embeddings? no

  • Tags: detection, coco, torch, transformers, detr, official

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("dfine-large-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

dfine-medium-coco-torch#

D-FINE Medium from D-FINE: Redefine Regression Task in DETRs as Fine-grained Distribution Refinement trained on COCO. Mid-size real-time object detector..

Details

  • Model name: dfine-medium-coco-torch

  • Model source: Peterande/D-FINE

  • Model author: Yansong Peng, et al.

  • Model license: Apache 2.0

  • Model size: 30.00 MB

  • Exposes embeddings? no

  • Tags: detection, coco, torch, transformers, detr, official

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("dfine-medium-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

dfine-nano-coco-torch#

D-FINE Nano from D-FINE: Redefine Regression Task in DETRs as Fine-grained Distribution Refinement trained on COCO. Ultra-lightweight real-time object detector..

Details

  • Model name: dfine-nano-coco-torch

  • Model source: Peterande/D-FINE

  • Model author: Yansong Peng, et al.

  • Model license: Apache 2.0

  • Model size: 14.00 MB

  • Exposes embeddings? no

  • Tags: detection, coco, torch, transformers, detr, official

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("dfine-nano-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

dfine-small-coco-torch#

D-FINE Small from D-FINE: Redefine Regression Task in DETRs as Fine-grained Distribution Refinement trained on COCO. Balanced real-time object detector..

Details

  • Model name: dfine-small-coco-torch

  • Model source: Peterande/D-FINE

  • Model author: Yansong Peng, et al.

  • Model license: Apache 2.0

  • Model size: 24.00 MB

  • Exposes embeddings? no

  • Tags: detection, coco, torch, transformers, detr, official

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("dfine-small-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

dfine-xlarge-coco-torch#

D-FINE XLarge from D-FINE: Redefine Regression Task in DETRs as Fine-grained Distribution Refinement trained on COCO. Achieves 55.8% AP at 78 FPS on T4 GPU..

Details

  • Model name: dfine-xlarge-coco-torch

  • Model source: Peterande/D-FINE

  • Model author: Yansong Peng, et al.

  • Model license: Apache 2.0

  • Model size: 123.00 MB

  • Exposes embeddings? no

  • Tags: detection, coco, torch, transformers, detr, official

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("dfine-xlarge-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

dinov2-vitb14-reg-torch#

Enhanced image search model that resists noise and errors for more reliable similarity matching.

Details

  • Model name: dinov2-vitb14-reg-torch

  • Model source: facebookresearch/dinov2

  • Model author: Maxime Oquab, et al.

  • Model license: Apache 2.0

  • Model size: 330.35 MB

  • Exposes embeddings? yes

  • Tags: embeddings, torch, dinov2, transformer, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("dinov2-vitb14-reg-torch")
13
14embeddings = dataset.compute_embeddings(model)

dinov2-vitb14-torch#

Creates searchable image fingerprints for finding similar pictures and organizing large photo collections.

Details

  • Model name: dinov2-vitb14-torch

  • Model source: facebookresearch/dinov2

  • Model author: Maxime Oquab, et al.

  • Model license: Apache 2.0

  • Model size: 330.33 MB

  • Exposes embeddings? yes

  • Tags: embeddings, torch, dinov2, transformer, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("dinov2-vitb14-torch")
13
14embeddings = dataset.compute_embeddings(model)

dinov2-vitg14-reg-torch#

Highest-capacity dinov2 search model with maximum stability for finding images across massive diverse datasets.

Details

  • Model name: dinov2-vitg14-reg-torch

  • Model source: facebookresearch/dinov2

  • Model author: Maxime Oquab, et al.

  • Model license: Apache 2.0

  • Model size: 4.23 GB

  • Exposes embeddings? yes

  • Tags: embeddings, torch, dinov2, transformer, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("dinov2-vitg14-reg-torch")
13
14embeddings = dataset.compute_embeddings(model)

dinov2-vitg14-torch#

Powerful image search engine that handles enormous photo collections with rich detail extraction.

Details

  • Model name: dinov2-vitg14-torch

  • Model source: facebookresearch/dinov2

  • Model author: Maxime Oquab, et al.

  • Model license: Apache 2.0

  • Model size: 4.23 GB

  • Exposes embeddings? yes

  • Tags: embeddings, torch, dinov2, transformer, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("dinov2-vitg14-torch")
13
14embeddings = dataset.compute_embeddings(model)

dinov2-vitl14-reg-torch#

Large stable model for finding and grouping similar images across big databases reliably.

Details

  • Model name: dinov2-vitl14-reg-torch

  • Model source: facebookresearch/dinov2

  • Model author: Maxime Oquab, et al.

  • Model license: Apache 2.0

  • Model size: 1.13 GB

  • Exposes embeddings? yes

  • Tags: embeddings, torch, dinov2, transformer, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("dinov2-vitl14-reg-torch")
13
14embeddings = dataset.compute_embeddings(model)

dinov2-vitl14-torch#

Large model that creates detailed image fingerprints for advanced search and automatic grouping.

Details

  • Model name: dinov2-vitl14-torch

  • Model source: facebookresearch/dinov2

  • Model author: Maxime Oquab, et al.

  • Model license: Apache 2.0

  • Model size: 1.13 GB

  • Exposes embeddings? yes

  • Tags: embeddings, torch, dinov2, transformer, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("dinov2-vitl14-torch")
13
14embeddings = dataset.compute_embeddings(model)

dinov2-vits14-reg-torch#

Compact stable model for image search that runs efficiently on phones and edge devices.

Details

  • Model name: dinov2-vits14-reg-torch

  • Model source: facebookresearch/dinov2

  • Model author: Maxime Oquab, et al.

  • Model license: Apache 2.0

  • Model size: 84.20 MB

  • Exposes embeddings? yes

  • Tags: embeddings, torch, dinov2, transformer, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("dinov2-vits14-reg-torch")
13
14embeddings = dataset.compute_embeddings(model)

dinov2-vits14-torch#

Small model enabling image search and similarity matching directly on mobile devices.

Details

  • Model name: dinov2-vits14-torch

  • Model source: facebookresearch/dinov2

  • Model author: Maxime Oquab, et al.

  • Model license: Apache 2.0

  • Model size: 84.19 MB

  • Exposes embeddings? yes

  • Tags: embeddings, torch, dinov2, transformer, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("dinov2-vits14-torch")
13
14embeddings = dataset.compute_embeddings(model)

efficientnet-b0-imagenet-torch#

Efficient image classifier optimized for mobile devices with excellent accuracy-efficiency tradeoff.

Details

  • Model name: efficientnet-b0-imagenet-torch

  • Model source: https://huggingface.co/google/efficientnet-b0

  • Model author: Mingxing Tan and Quoc V. Le

  • Model license: Apache 2.0

  • Model size: 81.80 MB

  • Exposes embeddings? no

  • Tags: classification, imagenet, torch, transformers, efficientnet, official

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("efficientnet-b0-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

efficientnet-b1-imagenet-torch#

Scaled efficient classifier with improved accuracy for slightly larger computational budgets.

Details

  • Model name: efficientnet-b1-imagenet-torch

  • Model source: https://huggingface.co/google/efficientnet-b1

  • Model author: Mingxing Tan and Quoc V. Le

  • Model license: Apache 2.0

  • Model size: 120.47 MB

  • Exposes embeddings? no

  • Tags: classification, imagenet, torch, transformers, efficientnet, official

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("efficientnet-b1-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

efficientnet-b2-imagenet-torch#

Balanced efficient model providing stronger performance while maintaining reasonable resource usage.

Details

  • Model name: efficientnet-b2-imagenet-torch

  • Model source: https://huggingface.co/google/efficientnet-b2

  • Model author: Mingxing Tan and Quoc V. Le

  • Model license: Apache 2.0

  • Model size: 140.63 MB

  • Exposes embeddings? no

  • Tags: classification, imagenet, torch, transformers, efficientnet, official

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("efficientnet-b2-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

efficientnet-b3-imagenet-torch#

Mid-scale efficient classifier delivering high accuracy for versatile deployment scenarios.

Details

  • Model name: efficientnet-b3-imagenet-torch

  • Model source: https://huggingface.co/google/efficientnet-b3

  • Model author: Mingxing Tan and Quoc V. Le

  • Model license: Apache 2.0

  • Model size: 188.64 MB

  • Exposes embeddings? no

  • Tags: classification, imagenet, torch, transformers, efficientnet, official

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("efficientnet-b3-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

efficientnet-b4-imagenet-torch#

Large efficient model with enhanced features for transfer learning applications.

Details

  • Model name: efficientnet-b4-imagenet-torch

  • Model source: https://huggingface.co/google/efficientnet-b4

  • Model author: Mingxing Tan and Quoc V. Le

  • Model license: Apache 2.0

  • Model size: 297.81 MB

  • Exposes embeddings? no

  • Tags: classification, imagenet, torch, transformers, efficientnet, official

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("efficientnet-b4-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

efficientnet-b5-imagenet-torch#

High-capacity efficient classifier prioritizing accuracy with available compute resources.

Details

  • Model name: efficientnet-b5-imagenet-torch

  • Model source: https://huggingface.co/google/efficientnet-b5

  • Model author: Mingxing Tan and Quoc V. Le

  • Model license: Apache 2.0

  • Model size: 467.25 MB

  • Exposes embeddings? no

  • Tags: classification, imagenet, torch, transformers, efficientnet, official

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("efficientnet-b5-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

efficientnet-b6-imagenet-torch#

Extended efficient model approaching state-of-the-art accuracy on challenging datasets.

Details

  • Model name: efficientnet-b6-imagenet-torch

  • Model source: https://huggingface.co/google/efficientnet-b6

  • Model author: Mingxing Tan and Quoc V. Le

  • Model license: Apache 2.0

  • Model size: 661.19 MB

  • Exposes embeddings? no

  • Tags: classification, imagenet, torch, transformers, efficientnet, official

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("efficientnet-b6-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

efficientnet-b7-imagenet-torch#

Maximum efficient classifier pushing performance boundaries while preserving efficiency principles.

Details

  • Model name: efficientnet-b7-imagenet-torch

  • Model source: https://huggingface.co/google/efficientnet-b7

  • Model author: Mingxing Tan and Quoc V. Le

  • Model license: Apache 2.0

  • Model size: 1018.36 MB

  • Exposes embeddings? no

  • Tags: classification, imagenet, torch, transformers, efficientnet, official

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("efficientnet-b7-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

faster-rcnn-resnet50-fpn-coco-torch#

Multi-scale object finder that accurately detects both small and large items in images.

Details

  • Model name: faster-rcnn-resnet50-fpn-coco-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Shaoqing Ren, et al.

  • Model license: BSD 3-Clause

  • Model size: 159.74 MB

  • Exposes embeddings? no

  • Tags: detection, coco, torch, faster-rcnn, resnet, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("faster-rcnn-resnet50-fpn-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

fcn-resnet101-coco-torch#

Creates detailed pixel-level labels for images, identifying and outlining twenty-one different object categories.

Details

  • Model name: fcn-resnet101-coco-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Jonathan Long, et al.

  • Model license: BSD 3-Clause

  • Model size: 207.71 MB

  • Exposes embeddings? no

  • Tags: segmentation, coco, torch, fcn, resnet, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("fcn-resnet101-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

fcn-resnet50-coco-torch#

Fast image labeler that quickly identifies and outlines objects for interactive editing and annotation.

Details

  • Model name: fcn-resnet50-coco-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Jonathan Long, et al.

  • Model license: BSD 3-Clause

  • Model size: 135.01 MB

  • Exposes embeddings? no

  • Tags: segmentation, coco, torch, fcn, resnet, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("fcn-resnet50-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

googlenet-imagenet-torch#

Classic image classifier providing reliable categorization and features for various computer vision projects..

Details

  • Model name: googlenet-imagenet-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Christian Szegedy, et al.

  • Model license: BSD 3-Clause

  • Model size: 49.73 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, torch, googlenet, official

Requirements

  • Packages: scipy, torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("googlenet-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

group-vit-segmentation-transformer-torch#

Hugging Face Transformers model for zero-shot semantic segmentation.

Details

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("group-vit-segmentation-transformer-torch",
13    text_prompt="A photo of a",
14    classes=["person", "dog", "cat", "bird", "car", "tree", "other"])
15
16dataset.apply_model(model, label_field="predictions")
17
18session = fo.launch_app(dataset)

inception-v3-imagenet-torch#

Efficient image classifier delivering accurate results with useful features for transfer learning applications.

Details

  • Model name: inception-v3-imagenet-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Christian Szegedy, et al.

  • Model license: BSD 3-Clause

  • Model size: 103.81 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, torch, inception, official

Requirements

  • Packages: scipy, torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("inception-v3-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

keypoint-rcnn-resnet50-fpn-coco-torch#

Finds people in images and maps their body joints for pose estimation and motion analysis.

Details

  • Model name: keypoint-rcnn-resnet50-fpn-coco-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Kaiming He, et al.

  • Model license: BSD 3-Clause

  • Model size: 226.05 MB

  • Exposes embeddings? no

  • Tags: keypoints, coco, torch, keypoint-rcnn, resnet, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("keypoint-rcnn-resnet50-fpn-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

mask-rcnn-resnet50-fpn-coco-torch#

Multi-scale object outliner using advanced architecture for accurate segmentation across different object sizes.

Details

  • Model name: mask-rcnn-resnet50-fpn-coco-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Kaiming He, et al.

  • Model license: BSD 3-Clause

  • Model size: 169.84 MB

  • Exposes embeddings? no

  • Tags: instances, coco, torch, mask-rcnn, resnet, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("mask-rcnn-resnet50-fpn-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

med-sam-2-video-torch#

Medical segmentation tool that outlines organs and structures in medical videos and 3D scans.

Details

  • Model name: med-sam-2-video-torch

  • Model source: MedicineToken/Medical-SAM2

  • Model author: Jiayuan Zhu, et al.

  • Model license: Apache 2.0

  • Model size: 74.46 MB

  • Exposes embeddings? no

  • Tags: segment-anything, torch, zero-shot, video, med-SAM, transformer, official

Requirements

  • Packages: torch, torchvision, sam2

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3from fiftyone import ViewField as F
 4from fiftyone.utils.huggingface import load_from_hub
 5
 6dataset = load_from_hub("Voxel51/BTCV-CT-as-video-MedSAM2-dataset")[:2]
 7
 8# Retaining detections from a single frame in the middle
 9# Note that SAM2 only propagates segmentation masks forward in a video
10(
11    dataset
12    .match_frames(F("frame_number") != 100)
13    .set_field("frames.gt_detections", None)
14    .save()
15)
16
17model = foz.load_zoo_model("med-sam-2-video-torch")
18
19# Segment inside boxes and propagate to all frames
20dataset.apply_model(
21    model,
22    label_field="pred_segmentations",
23    prompt_field="frames.gt_detections",
24)
25
26session = fo.launch_app(dataset)

mnasnet0.5-imagenet-torch#

Ultra-lightweight image classifier designed by AI for running directly on phones and IoT devices.

Details

  • Model name: mnasnet0.5-imagenet-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Mingxing Tan, et al.

  • Model license: BSD 3-Clause

  • Model size: 8.59 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, torch, mnasnet, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("mnasnet0.5-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

mnasnet1.0-imagenet-torch#

Mobile-optimized classifier balancing size and accuracy for efficient on-device image recognition.

Details

  • Model name: mnasnet1.0-imagenet-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Mingxing Tan, et al.

  • Model license: BSD 3-Clause

  • Model size: 16.92 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, torch, mnasnet, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("mnasnet1.0-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

mobilenet-v2-imagenet-torch#

Mobile-friendly image classifier optimized for quick training and deployment on resource-limited devices.

Details

  • Model name: mobilenet-v2-imagenet-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Mark Sandler, et al.

  • Model license: BSD 3-Clause

  • Model size: 13.55 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, torch, mobilenet, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("mobilenet-v2-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

omdet-turbo-swin-tiny-torch#

Real-time detector that finds any object you describe in words, perfect for live video analysis.

Details

Requirements

  • Packages: torch, torchvision, transformers>=4.51

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12
13classes = ["person", "dog", "cat", "bird", "car", "tree", "chair"]
14
15model = foz.load_zoo_model(
16    "omdet-turbo-swin-tiny-torch",
17    classes=classes,
18)
19
20dataset.apply_model(model, label_field="predictions")
21
22session = fo.launch_app(dataset)

open-clip-torch#

Connects images with text descriptions enabling search by words and automatic content filtering systems.

Details

  • Model name: open-clip-torch

  • Model source: mlfoundations/open_clip

  • Model author: Gabriel Ilharco, et al.

  • Model license: MIT

  • Exposes embeddings? yes

  • Tags: classification, logits, embeddings, torch, clip, zero-shot, transformer

Requirements

  • Packages: torch, torchvision, open_clip_torch

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("open-clip-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
17
18#
19# Make zero-shot predictions with custom classes
20#
21
22model = foz.load_zoo_model(
23    "open-clip-torch",
24    text_prompt="A photo of a",
25    classes=["person", "dog", "cat", "bird", "car", "tree", "chair"],
26)
27
28dataset.apply_model(model, label_field="predictions")
29session.refresh()

owlvit-base-patch16-torch#

Finds any object you name in pictures using 16x16 image patches without needing specific training for those items.

Details

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12
13classes = ["person", "dog", "cat", "bird", "car", "tree", "chair"]
14
15model = foz.load_zoo_model(
16    "owlvit-base-patch16-torch",
17    classes=classes,
18)
19
20dataset.apply_model(model, label_field="predictions")
21
22session = fo.launch_app(dataset)

owlvit-base-patch32-torch#

Finds any object you name in pictures using efficient 32x32 image patches without needing specific training.

Details

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12
13classes = ["person", "dog", "cat", "bird", "car", "tree", "chair"]
14
15model = foz.load_zoo_model(
16    "owlvit-base-patch32-torch",
17    classes=classes,
18)
19
20dataset.apply_model(model, label_field="predictions")
21
22session = fo.launch_app(dataset)

owlvit-large-patch14-torch#

Large OWL-ViT zero-shot object detector with ViT-L/14 backbone. Achieves higher accuracy than base models, especially for smaller objects..

Details

  • Model name: owlvit-large-patch14-torch

  • Model source: https://huggingface.co/google/owlvit-large-patch14

  • Model author: Matthias Minderer, et al.

  • Model license: Apache 2.0

  • Model size: 1.62 GB

  • Exposes embeddings? yes

  • Tags: detection, logits, embeddings, torch, transformers, zero-shot, official

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12
13classes = ["person", "dog", "cat", "bird", "car", "tree", "chair"]
14
15model = foz.load_zoo_model(
16    "owlvit-large-patch14-torch",
17    classes=classes,
18)
19
20dataset.apply_model(model, label_field="predictions")
21
22session = fo.launch_app(dataset)

resnet101-imagenet-torch#

Deep image recognition model delivering high accuracy for demanding classification and analysis tasks.

Details

  • Model name: resnet101-imagenet-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Kaiming He, et al.

  • Model license: BSD 3-Clause

  • Model size: 170.45 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, torch, resnet, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("resnet101-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

resnet152-imagenet-torch#

Very deep classifier providing the richest visual features for precision-critical image understanding applications.

Details

  • Model name: resnet152-imagenet-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Kaiming He, et al.

  • Model license: BSD 3-Clause

  • Model size: 230.34 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, torch, resnet, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("resnet152-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

resnet18-imagenet-torch#

Lightweight image classifier designed for fast recognition on phones and other resource-limited devices.

Details

  • Model name: resnet18-imagenet-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Kaiming He, et al.

  • Model license: BSD 3-Clause

  • Model size: 44.66 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, torch, resnet, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("resnet18-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

resnet34-imagenet-torch#

Balanced image classifier offering good accuracy and speed for everyday computer vision needs.

Details

  • Model name: resnet34-imagenet-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Kaiming He, et al.

  • Model license: BSD 3-Clause

  • Model size: 83.26 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, torch, resnet, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("resnet34-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

resnet50-imagenet-torch#

Most popular image recognition backbone widely used as starting point for custom vision projects.

Details

  • Model name: resnet50-imagenet-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Kaiming He, et al.

  • Model license: BSD 3-Clause

  • Model size: 97.75 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, torch, resnet, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("resnet50-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

resnext101-32x8d-imagenet-torch#

Powerful image classifier with enhanced capacity for handling complex visual recognition challenges effectively.

Details

  • Model name: resnext101-32x8d-imagenet-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Saining Xie, et al.

  • Model license: BSD 3-Clause

  • Model size: 339.59 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, torch, resnext, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("resnext101-32x8d-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

resnext50-32x4d-imagenet-torch#

Efficient advanced classifier delivering strong accuracy with reasonable computing requirements for practical deployments.

Details

  • Model name: resnext50-32x4d-imagenet-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Saining Xie, et al.

  • Model license: BSD 3-Clause

  • Model size: 95.79 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, torch, resnext

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("resnext50-32x4d-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

retinanet-resnet50-fpn-coco-torch#

Fast object detector that quickly finds and boxes eighty common items in any image.

Details

  • Model name: retinanet-resnet50-fpn-coco-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Tsung-Yi Lin, et al.

  • Model license: BSD 3-Clause

  • Model size: 130.27 MB

  • Exposes embeddings? no

  • Tags: detection, coco, torch, retinanet, resnet

Requirements

  • Packages: torch, torchvision>=0.8.0

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("retinanet-resnet50-fpn-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

rtdetr-l-coco-torch#

Modern real-time object detector that finds items without complex post-processing for responsive applications.

Details

  • Model name: rtdetr-l-coco-torch

  • Model source: https://docs.ultralytics.com/models/rtdetr/

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 63.43 MB

  • Exposes embeddings? no

  • Tags: detection, coco, torch, transformer, rtdetr, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.2.0

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("rtdetr-l-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

rtdetr-v2-m-coco-torch#

Balanced real-time object detector offering improved accuracy for production use.

Details

  • Model name: rtdetr-v2-m-coco-torch

  • Model source: https://huggingface.co/PekingU/rtdetr_v2_r50vd

  • Model author: Wenyu Lv, et al.

  • Model license: Apache-2.0

  • Model size: 328.41 MB

  • Exposes embeddings? no

  • Tags: detection, coco, torch, transformers, rtdetr, official

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("rtdetr-v2-m-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

rtdetr-v2-s-coco-torch#

Lightweight real-time object detector optimized for speed on edge devices.

Details

  • Model name: rtdetr-v2-s-coco-torch

  • Model source: https://huggingface.co/PekingU/rtdetr_v2_r18vd

  • Model author: Wenyu Lv, et al.

  • Model license: Apache-2.0

  • Model size: 154.32 MB

  • Exposes embeddings? no

  • Tags: detection, coco, torch, transformers, rtdetr

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("rtdetr-v2-s-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

rtdetr-x-coco-torch#

High-capacity object detector delivering very precise results at speeds suitable for production use.

Details

  • Model name: rtdetr-x-coco-torch

  • Model source: https://docs.ultralytics.com/models/rtdetr/

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 129.47 MB

  • Exposes embeddings? no

  • Tags: detection, coco, torch, transformer, rtdetr, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.2.0

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("rtdetr-x-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

segformer-b0-ade20k-torch#

Efficient transformer-based semantic segmentation model for scene parsing with 150 classes.

Details

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("segformer-b0-ade20k-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

segformer-b1-ade20k-torch#

Balanced SegFormer model providing good accuracy-efficiency tradeoff for scene understanding.

Details

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("segformer-b1-ade20k-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

segformer-b2-ade20k-torch#

Medium-sized SegFormer delivering enhanced segmentation quality for complex scenes.

Details

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("segformer-b2-ade20k-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

segformer-b3-ade20k-torch#

Larger SegFormer model with improved accuracy for detailed semantic segmentation.

Details

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("segformer-b3-ade20k-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

segformer-b4-ade20k-torch#

High-capacity SegFormer achieving excellent results on challenging segmentation tasks.

Details

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("segformer-b4-ade20k-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

segformer-b5-ade20k-torch#

Largest SegFormer model delivering the best semantic segmentation performance in its family.

Details

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("segformer-b5-ade20k-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

segment-anything-2-hiera-base-plus-image-torch#

Accurate image segmentation model for editing, labeling, and creative work with still pictures.

Details

  • Model name: segment-anything-2-hiera-base-plus-image-torch

  • Model source: https://ai.meta.com/sam2/

  • Model author: Nikhila Ravi, et al.

  • Model license: Apache 2.0, BSD 3-Clause

  • Model size: 308.51 MB

  • Exposes embeddings? no

  • Tags: segment-anything, torch, zero-shot, transformer, official

Requirements

  • Packages: torch, torchvision, sam2

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("segment-anything-2-hiera-base-plus-image-torch")
13
14# Segment inside boxes
15dataset.apply_model(
16    model,
17    label_field="segmentations",
18    prompt_field="ground_truth",  # can contain Detections or Keypoints
19)
20
21# Full automatic segmentations
22dataset.apply_model(model, label_field="auto")
23
24session = fo.launch_app(dataset)

segment-anything-2-hiera-base-plus-video-torch#

Video segmentation model that tracks and outlines objects throughout clips for editing and analysis.

Details

  • Model name: segment-anything-2-hiera-base-plus-video-torch

  • Model source: https://ai.meta.com/sam2/

  • Model author: Nikhila Ravi, et al.

  • Model license: Apache 2.0, BSD 3-Clause

  • Model size: 308.51 MB

  • Exposes embeddings? no

  • Tags: segment-anything, torch, zero-shot, video, transformer, official

Requirements

  • Packages: torch, torchvision, sam2

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3from fiftyone import ViewField as F
 4
 5dataset = foz.load_zoo_dataset("quickstart-video", max_samples=2)
 6
 7# Only retain detections in the first frame
 8(
 9    dataset
10    .match_frames(F("frame_number") > 1)
11    .set_field("frames.detections", None)
12    .save()
13)
14
15model = foz.load_zoo_model("segment-anything-2-hiera-base-plus-video-torch")
16
17# Segment inside boxes and propagate to all frames
18dataset.apply_model(
19    model,
20    label_field="segmentations",
21    prompt_field="frames.detections",  # can contain Detections or Keypoints
22)
23
24session = fo.launch_app(dataset)

segment-anything-2-hiera-large-image-torch#

High-quality image segmenter producing detailed masks for demanding professional editing and annotation tasks.

Details

  • Model name: segment-anything-2-hiera-large-image-torch

  • Model source: https://ai.meta.com/sam2/

  • Model author: Nikhila Ravi, et al.

  • Model license: Apache 2.0, BSD 3-Clause

  • Model size: 856.35 MB

  • Exposes embeddings? no

  • Tags: segment-anything, torch, zero-shot, transformer, official

Requirements

  • Packages: torch, torchvision, sam2

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("segment-anything-2-hiera-large-image-torch")
13
14# Segment inside boxes
15dataset.apply_model(
16    model,
17    label_field="segmentations",
18    prompt_field="ground_truth",  # can contain Detections or Keypoints
19)
20
21# Full automatic segmentations
22dataset.apply_model(model, label_field="auto")
23
24session = fo.launch_app(dataset)

segment-anything-2-hiera-large-video-torch#

Advanced video segmenter providing fine object tracking throughout full videos for post-production work.

Details

  • Model name: segment-anything-2-hiera-large-video-torch

  • Model source: https://ai.meta.com/sam2/

  • Model author: Nikhila Ravi, et al.

  • Model license: Apache 2.0, BSD 3-Clause

  • Model size: 856.35 MB

  • Exposes embeddings? no

  • Tags: segment-anything, torch, zero-shot, video, transformer

Requirements

  • Packages: torch, torchvision, sam2

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3from fiftyone import ViewField as F
 4
 5dataset = foz.load_zoo_dataset("quickstart-video", max_samples=2)
 6
 7# Only retain detections in the first frame
 8(
 9    dataset
10    .match_frames(F("frame_number") > 1)
11    .set_field("frames.detections", None)
12    .save()
13)
14
15model = foz.load_zoo_model("segment-anything-2-hiera-large-video-torch")
16
17# Segment inside boxes and propagate to all frames
18dataset.apply_model(
19    model,
20    label_field="segmentations",
21    prompt_field="frames.detections",  # can contain Detections or Keypoints
22)
23
24session = fo.launch_app(dataset)

segment-anything-2-hiera-small-image-torch#

Fast image segmentation model that runs efficiently on laptops and edge computing devices.

Details

  • Model name: segment-anything-2-hiera-small-image-torch

  • Model source: https://ai.meta.com/sam2/

  • Model author: Nikhila Ravi, et al.

  • Model license: Apache 2.0, BSD 3-Clause

  • Model size: 175.77 MB

  • Exposes embeddings? no

  • Tags: segment-anything, torch, zero-shot, transformer, official

Requirements

  • Packages: torch, torchvision, sam2

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("segment-anything-2-hiera-small-image-torch")
13
14# Segment inside boxes
15dataset.apply_model(
16    model,
17    label_field="segmentations",
18    prompt_field="ground_truth",  # can contain Detections or Keypoints
19)
20
21# Full automatic segmentations
22dataset.apply_model(model, label_field="auto")
23
24session = fo.launch_app(dataset)

segment-anything-2-hiera-small-video-torch#

Quick video segmentation model delivering rapid object tracking on standard graphics cards.

Details

  • Model name: segment-anything-2-hiera-small-video-torch

  • Model source: https://ai.meta.com/sam2/

  • Model author: Nikhila Ravi, et al.

  • Model license: Apache 2.0, BSD 3-Clause

  • Model size: 175.77 MB

  • Exposes embeddings? no

  • Tags: segment-anything, torch, zero-shot, video, transformer, official

Requirements

  • Packages: torch, torchvision, sam2

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3from fiftyone import ViewField as F
 4
 5dataset = foz.load_zoo_dataset("quickstart-video", max_samples=2)
 6
 7# Only retain detections in the first frame
 8(
 9    dataset
10    .match_frames(F("frame_number") > 1)
11    .set_field("frames.detections", None)
12    .save()
13)
14
15model = foz.load_zoo_model("segment-anything-2-hiera-small-video-torch")
16
17# Segment inside boxes and propagate to all frames
18dataset.apply_model(
19    model,
20    label_field="segmentations",
21    prompt_field="frames.detections",  # can contain Detections or Keypoints
22)
23
24session = fo.launch_app(dataset)

segment-anything-2-hiera-tiny-image-torch#

Smallest image segmentation model offering instant results for mobile apps and embedded systems.

Details

  • Model name: segment-anything-2-hiera-tiny-image-torch

  • Model source: https://ai.meta.com/sam2/

  • Model author: Nikhila Ravi, et al.

  • Model license: Apache 2.0, BSD 3-Clause

  • Model size: 148.68 MB

  • Exposes embeddings? no

  • Tags: segment-anything, torch, zero-shot

Requirements

  • Packages: torch, torchvision, sam2

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("segment-anything-2-hiera-tiny-image-torch")
13
14# Segment inside boxes
15dataset.apply_model(
16    model,
17    label_field="segmentations",
18    prompt_field="ground_truth",  # can contain Detections or Keypoints
19)
20
21# Full automatic segmentations
22dataset.apply_model(model, label_field="auto")
23
24session = fo.launch_app(dataset)

segment-anything-2-hiera-tiny-video-torch#

Tiny video segmenter enabling real-time object tracking on phones and compact devices.

Details

  • Model name: segment-anything-2-hiera-tiny-video-torch

  • Model source: https://ai.meta.com/sam2/

  • Model author: Nikhila Ravi, et al.

  • Model license: Apache 2.0, BSD 3-Clause

  • Model size: 148.68 MB

  • Exposes embeddings? no

  • Tags: segment-anything, torch, zero-shot, video, transformer, official

Requirements

  • Packages: torch, torchvision, sam2

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3from fiftyone import ViewField as F
 4
 5dataset = foz.load_zoo_dataset("quickstart-video", max_samples=2)
 6
 7# Only retain detections in the first frame
 8(
 9    dataset
10    .match_frames(F("frame_number") > 1)
11    .set_field("frames.detections", None)
12    .save()
13)
14
15model = foz.load_zoo_model("segment-anything-2-hiera-tiny-video-torch")
16
17# Segment inside boxes and propagate to all frames
18dataset.apply_model(
19    model,
20    label_field="segmentations",
21    prompt_field="frames.detections",  # can contain Detections or Keypoints
22)
23
24session = fo.launch_app(dataset)

segment-anything-2.1-hiera-base-plus-image-torch#

Updated image segmenter with improved mask accuracy for everyday editing and dataset creation.

Details

  • Model name: segment-anything-2.1-hiera-base-plus-image-torch

  • Model source: https://ai.meta.com/sam2/

  • Model author: Nikhila Ravi, et al.

  • Model license: Apache 2.0, BSD 3-Clause

  • Model size: 308.62 MB

  • Exposes embeddings? no

  • Tags: segment-anything, torch, zero-shot, transformer, official

Requirements

  • Packages: torch, torchvision, sam2

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("segment-anything-2.1-hiera-base-plus-image-torch")
13
14# Segment inside boxes
15dataset.apply_model(
16    model,
17    label_field="segmentations",
18    prompt_field="ground_truth",  # can contain Detections or Keypoints
19)
20
21# Full automatic segmentations
22dataset.apply_model(model, label_field="auto")
23
24session = fo.launch_app(dataset)

segment-anything-2.1-hiera-base-plus-video-torch#

Enhanced video segmenter with better tracking quality for video analysis and scene understanding.

Details

  • Model name: segment-anything-2.1-hiera-base-plus-video-torch

  • Model source: https://ai.meta.com/sam2/

  • Model author: Nikhila Ravi, et al.

  • Model license: Apache 2.0, BSD 3-Clause

  • Model size: 308.62 MB

  • Exposes embeddings? no

  • Tags: segment-anything, torch, zero-shot, video, transformer, official

Requirements

  • Packages: torch, torchvision, sam2

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3from fiftyone import ViewField as F
 4
 5dataset = foz.load_zoo_dataset("quickstart-video", max_samples=2)
 6
 7# Only retain detections in the first frame
 8(
 9    dataset
10    .match_frames(F("frame_number") > 1)
11    .set_field("frames.detections", None)
12    .save()
13)
14
15model = foz.load_zoo_model("segment-anything-2.1-hiera-base-plus-video-torch")
16
17# Segment inside boxes and propagate to all frames
18dataset.apply_model(
19    model,
20    label_field="segmentations",
21    prompt_field="frames.detections",  # can contain Detections or Keypoints
22)
23
24session = fo.launch_app(dataset)

segment-anything-2.1-hiera-large-image-torch#

Large updated model offering even finer masks for high-resolution professional image workflows.

Details

  • Model name: segment-anything-2.1-hiera-large-image-torch

  • Model source: https://ai.meta.com/sam2/

  • Model author: Nikhila Ravi, et al.

  • Model license: Apache 2.0, BSD 3-Clause

  • Model size: 856.48 MB

  • Exposes embeddings? no

  • Tags: segment-anything, torch, zero-shot, transformer, official

Requirements

  • Packages: torch, torchvision, sam2

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("segment-anything-2.1-hiera-large-image-torch")
13
14# Segment inside boxes
15dataset.apply_model(
16    model,
17    label_field="segmentations",
18    prompt_field="ground_truth",  # can contain Detections or Keypoints
19)
20
21# Full automatic segmentations
22dataset.apply_model(model, label_field="auto")
23
24session = fo.launch_app(dataset)

segment-anything-2.1-hiera-large-video-torch#

Large video model producing exceptionally detailed masks throughout long videos for intensive production.

Details

  • Model name: segment-anything-2.1-hiera-large-video-torch

  • Model source: https://ai.meta.com/sam2/

  • Model author: Nikhila Ravi, et al.

  • Model license: Apache 2.0, BSD 3-Clause

  • Model size: 856.48 MB

  • Exposes embeddings? no

  • Tags: segment-anything, torch, zero-shot, video, transformer, official

Requirements

  • Packages: torch, torchvision, sam2

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3from fiftyone import ViewField as F
 4
 5dataset = foz.load_zoo_dataset("quickstart-video", max_samples=2)
 6
 7# Only retain detections in the first frame
 8(
 9    dataset
10    .match_frames(F("frame_number") > 1)
11    .set_field("frames.detections", None)
12    .save()
13)
14
15model = foz.load_zoo_model("segment-anything-2.1-hiera-large-video-torch")
16
17# Segment inside boxes and propagate to all frames
18dataset.apply_model(
19    model,
20    label_field="segmentations",
21    prompt_field="frames.detections",  # can contain Detections or Keypoints
22)
23
24session = fo.launch_app(dataset)

segment-anything-2.1-hiera-small-image-torch#

Balanced updated segmenter combining speed and accuracy for edge device image processing.

Details

  • Model name: segment-anything-2.1-hiera-small-image-torch

  • Model source: https://ai.meta.com/sam2/

  • Model author: Nikhila Ravi, et al.

  • Model license: Apache 2.0, BSD 3-Clause

  • Model size: 175.87 MB

  • Exposes embeddings? no

  • Tags: segment-anything, torch, zero-shot, transformer, official

Requirements

  • Packages: torch, torchvision, sam2

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("segment-anything-2.1-hiera-small-image-torch")
13
14# Segment inside boxes
15dataset.apply_model(
16    model,
17    label_field="segmentations",
18    prompt_field="ground_truth",  # can contain Detections or Keypoints
19)
20
21# Full automatic segmentations
22dataset.apply_model(model, label_field="auto")
23
24session = fo.launch_app(dataset)

segment-anything-2.1-hiera-small-video-torch#

Improved video segmenter maintaining quick performance on compact hardware while enhancing mask quality.

Details

  • Model name: segment-anything-2.1-hiera-small-video-torch

  • Model source: https://ai.meta.com/sam2/

  • Model author: Nikhila Ravi, et al.

  • Model license: Apache 2.0, BSD 3-Clause

  • Model size: 175.87 MB

  • Exposes embeddings? no

  • Tags: segment-anything, torch, zero-shot, video, transformer, official

Requirements

  • Packages: torch, torchvision, sam2

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3from fiftyone import ViewField as F
 4
 5dataset = foz.load_zoo_dataset("quickstart-video", max_samples=2)
 6
 7# Only retain detections in the first frame
 8(
 9    dataset
10    .match_frames(F("frame_number") > 1)
11    .set_field("frames.detections", None)
12    .save()
13)
14
15model = foz.load_zoo_model("segment-anything-2.1-hiera-small-video-torch")
16
17# Segment inside boxes and propagate to all frames
18dataset.apply_model(
19    model,
20    label_field="segmentations",
21    prompt_field="frames.detections",  # can contain Detections or Keypoints
22)
23
24session = fo.launch_app(dataset)

segment-anything-2.1-hiera-tiny-image-torch#

Enhanced mobile image segmenter for apps, augmented reality filters, and on-device processing.

Details

  • Model name: segment-anything-2.1-hiera-tiny-image-torch

  • Model source: https://ai.meta.com/sam2/

  • Model author: Nikhila Ravi, et al.

  • Model license: Apache 2.0, BSD 3-Clause

  • Model size: 148.68 MB

  • Exposes embeddings? no

  • Tags: segment-anything, torch, zero-shot, transformer

Requirements

  • Packages: torch, torchvision, sam2

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("segment-anything-2.1-hiera-tiny-image-torch")
13
14# Segment inside boxes
15dataset.apply_model(
16    model,
17    label_field="segmentations",
18    prompt_field="ground_truth",  # can contain Detections or Keypoints
19)
20
21# Full automatic segmentations
22dataset.apply_model(model, label_field="auto")
23
24session = fo.launch_app(dataset)

segment-anything-2.1-hiera-tiny-video-torch#

Upgraded mobile video segmenter for live effects on phones, wearables, and smart cameras.

Details

  • Model name: segment-anything-2.1-hiera-tiny-video-torch

  • Model source: https://ai.meta.com/sam2/

  • Model author: Nikhila Ravi, et al.

  • Model license: Apache 2.0, BSD 3-Clause

  • Model size: 148.68 MB

  • Exposes embeddings? no

  • Tags: segment-anything, torch, zero-shot, video, transformer

Requirements

  • Packages: torch, torchvision, sam2

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3from fiftyone import ViewField as F
 4
 5dataset = foz.load_zoo_dataset("quickstart-video", max_samples=2)
 6
 7# Only retain detections in the first frame
 8(
 9    dataset
10    .match_frames(F("frame_number") > 1)
11    .set_field("frames.detections", None)
12    .save()
13)
14
15model = foz.load_zoo_model("segment-anything-2.1-hiera-tiny-video-torch")
16
17# Segment inside boxes and propagate to all frames
18dataset.apply_model(
19    model,
20    label_field="segmentations",
21    prompt_field="frames.detections",  # can contain Detections or Keypoints
22)
23
24session = fo.launch_app(dataset)

segment-anything-vitb-torch#

Interactive segmentation tool that instantly outlines any object you point to or describe.

Details

  • Model name: segment-anything-vitb-torch

  • Model source: https://segment-anything.com

  • Model author: Alexander Kirillov, et al.

  • Model license: Apache 2.0

  • Model size: 357.67 MB

  • Exposes embeddings? no

  • Tags: segment-anything, sa-1b, torch, zero-shot, transformer, official

Requirements

  • Packages: torch, torchvision, segment-anything

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("segment-anything-vitb-torch")
13
14# Segment inside boxes
15dataset.apply_model(
16    model,
17    label_field="segmentations",
18    prompt_field="ground_truth",  # can contain Detections or Keypoints
19)
20
21# Full automatic segmentations
22dataset.apply_model(model, label_field="auto")
23
24session = fo.launch_app(dataset)

segment-anything-vith-torch#

Highest quality segmentation model creating extremely detailed masks for research and large-scale annotation projects.

Details

  • Model name: segment-anything-vith-torch

  • Model source: https://segment-anything.com

  • Model author: Alexander Kirillov, et al.

  • Model license: Apache 2.0

  • Model size: 2.39 GB

  • Exposes embeddings? no

  • Tags: segment-anything, sa-1b, torch, zero-shot, transformer, official

Requirements

  • Packages: torch, torchvision, segment-anything

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("segment-anything-vith-torch")
13
14# Segment inside boxes
15dataset.apply_model(
16    model,
17    label_field="segmentations",
18    prompt_field="ground_truth",  # can contain Detections or Keypoints
19)
20
21# Full automatic segmentations
22dataset.apply_model(model, label_field="auto")
23
24session = fo.launch_app(dataset)

segment-anything-vitl-torch#

Large segmentation model producing finer object outlines for professional editing and labeling workflows.

Details

  • Model name: segment-anything-vitl-torch

  • Model source: https://segment-anything.com

  • Model author: Alexander Kirillov, et al.

  • Model license: Apache 2.0

  • Model size: 1.16 GB

  • Exposes embeddings? no

  • Tags: segment-anything, sa-1b, torch, zero-shot, transformer

Requirements

  • Packages: torch, torchvision, segment-anything

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("segment-anything-vitl-torch")
13
14# Segment inside boxes
15dataset.apply_model(
16    model,
17    label_field="segmentations",
18    prompt_field="ground_truth",  # can contain Detections or Keypoints
19)
20
21# Full automatic segmentations
22dataset.apply_model(model, label_field="auto")
23
24session = fo.launch_app(dataset)

segmentation-transformer-torch#

Hugging Face Transformers model for semantic segmentation.

Details

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("segmentation-transformer-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

shufflenetv2-0.5x-imagenet-torch#

Ultra-small image classifier for tiny devices with very limited power and memory.

Details

  • Model name: shufflenetv2-0.5x-imagenet-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Ningning Ma, et al.

  • Model license: BSD 3-Clause

  • Model size: 5.28 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, torch, shufflenet, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("shufflenetv2-0.5x-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

shufflenetv2-1.0x-imagenet-torch#

Mobile image classifier that works efficiently on phones with modest computing resources.

Details

  • Model name: shufflenetv2-1.0x-imagenet-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Ningning Ma, et al.

  • Model license: BSD 3-Clause

  • Model size: 8.79 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, torch, shufflenet, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("shufflenetv2-1.0x-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

siglip-base-patch16-224-torch#

Hugging Face Transformers model for zero-shot image classification.

Details

Requirements

  • Packages: torch, torchvision, transformers>=4.51

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12
13classes = ["person", "dog", "cat", "bird", "car", "tree", "chair"]
14
15model = foz.load_zoo_model(
16    "siglip-base-patch16-224-torch",
17    classes=classes,
18)
19
20dataset.apply_model(model, label_field="predictions")
21
22session = fo.launch_app(dataset)

squeezenet-1.1-imagenet-torch#

Tiny image classifier that fits in just five megabytes for embedded devices.

Details

  • Model name: squeezenet-1.1-imagenet-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Forrest Iandola

  • Model license: BSD 2-Clause

  • Model size: 4.74 MB

  • Exposes embeddings? no

  • Tags: classification, imagenet, torch, squeezenet

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("squeezenet-1.1-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

squeezenet-imagenet-torch#

Ultra-compact image classifier perfect for severely resource-constrained hardware and applications.

Details

  • Model name: squeezenet-imagenet-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Forrest Iandola

  • Model license: BSD 2-Clause

  • Model size: 4.79 MB

  • Exposes embeddings? no

  • Tags: classification, imagenet, torch, squeezenet, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("squeezenet-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

swin-v2-base-torch#

Base hierarchical transformer delivering strong results across vision tasks.

Details

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("swin-v2-base-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

swin-v2-large-torch#

Large hierarchical transformer with enhanced capacity for demanding applications.

Details

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("swin-v2-large-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

swin-v2-small-torch#

Small hierarchical transformer balancing efficiency and performance for practical use.

Details

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("swin-v2-small-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

swin-v2-tiny-torch#

Tiny hierarchical transformer for efficient visual recognition on edge devices.

Details

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("swin-v2-tiny-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

vgg11-bn-imagenet-torch#

Classic image classifier with stable training useful for various computer vision projects.

Details

  • Model name: vgg11-bn-imagenet-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Karen Simonyan, et al.

  • Model license: BSD 3-Clause

  • Model size: 506.88 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, torch, vgg, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("vgg11-bn-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

vgg11-imagenet-torch#

Simple baseline image classifier valuable for research experimentation and learning purposes.

Details

  • Model name: vgg11-imagenet-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Karen Simonyan, et al.

  • Model license: BSD 3-Clause

  • Model size: 506.84 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, torch, vgg, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("vgg11-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

vgg13-bn-imagenet-torch#

Deeper classic classifier providing stable training process and solid accuracy results overall.

Details

  • Model name: vgg13-bn-imagenet-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Karen Simonyan, et al.

  • Model license: BSD 3-Clause

  • Model size: 507.59 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, torch, vgg, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("vgg13-bn-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

vgg13-imagenet-torch#

Straightforward image classifier valued for easy experimentation and model compression studies.

Details

  • Model name: vgg13-imagenet-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Karen Simonyan, et al.

  • Model license: BSD 3-Clause

  • Model size: 507.54 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, torch, vgg, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("vgg13-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

vgg16-bn-imagenet-torch#

Popular feature extractor widely used for detection, style transfer, and medical imaging.

Details

  • Model name: vgg16-bn-imagenet-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Karen Simonyan, et al.

  • Model license: BSD 3-Clause

  • Model size: 527.87 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, torch, vgg, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("vgg16-bn-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

vgg16-imagenet-torch#

PyTorch version of the popular classifier ready for modern deep learning workflows.

Details

  • Model name: vgg16-imagenet-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Karen Simonyan, et al.

  • Model license: BSD 3-Clause

  • Model size: 527.80 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, torch, vgg, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("vgg16-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

vgg19-bn-imagenet-torch#

Deep classic model providing rich features for style transfer and interpretability analysis.

Details

  • Model name: vgg19-bn-imagenet-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Karen Simonyan, et al.

  • Model license: BSD 3-Clause

  • Model size: 548.14 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, torch, vgg, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("vgg19-bn-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

vgg19-imagenet-torch#

Deep image classifier delivering detailed features for creative applications and research projects.

Details

  • Model name: vgg19-imagenet-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Karen Simonyan, et al.

  • Model license: BSD 3-Clause

  • Model size: 548.05 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, torch, vgg, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("vgg19-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

vit-base-patch16-224-imagenet-torch#

Modern image classifier that recognizes objects and provides useful features for various computer vision tasks.

Details

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("vit-base-patch16-224-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

wide-resnet101-2-imagenet-torch#

Extra-wide deep classifier for high-precision image recognition and advanced transfer learning tasks.

Details

  • Model name: wide-resnet101-2-imagenet-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Sergey Zagoruyko, et al.

  • Model license: BSD 3-Clause

  • Model size: 242.90 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, torch, wide-resnet, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("wide-resnet101-2-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

wide-resnet50-2-imagenet-torch#

Wide classifier offering stronger accuracy and better features for adapting to new tasks.

Details

  • Model name: wide-resnet50-2-imagenet-torch

  • Model source: https://pytorch.org/vision/main/models.html

  • Model author: Sergey Zagoruyko, et al.

  • Model license: BSD 3-Clause

  • Model size: 131.82 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, torch, wide-resnet, official

Requirements

  • Packages: torch, torchvision

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("wide-resnet50-2-imagenet-torch")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

yolo-nas-torch#

AI-designed detector family offering three model variants for diverse deployment scenarios.

Details

  • Model name: yolo-nas-torch

  • Model source: Deci-AI/super-gradients

  • Model author: Shay Aharon, et al.

  • Model license: Apache 2.0

  • Exposes embeddings? no

  • Tags: detection, torch, yolo, official

Requirements

  • Packages: torch, torchvision, super-gradients

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolo-nas-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolo11l-coco-torch#

Real-time object detector balancing high accuracy with fast processing speeds effectively.

Details

  • Model name: yolo11l-coco-torch

  • Model source: https://docs.ultralytics.com/models/yolov11/

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 49.01 MB

  • Exposes embeddings? no

  • Tags: detection, coco, torch, yolo, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.3.0

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolo11l-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolo11l-seg-coco-torch#

Model creating detailed object outlines for precise image editing and analysis.

Details

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.3.0

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolo11l-seg-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolo11m-coco-torch#

Object detector offering good balance between speed and accuracy for most applications.

Details

  • Model name: yolo11m-coco-torch

  • Model source: https://docs.ultralytics.com/models/yolov11/

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 38.80 MB

  • Exposes embeddings? no

  • Tags: detection, coco, torch, yolo, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.3.0

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolo11m-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolo11m-seg-coco-torch#

Model generating object masks efficiently for everyday segmentation tasks.

Details

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.3.0

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolo11m-seg-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolo11n-coco-torch#

Object detector designed specifically for phones and other edge computing devices.

Details

  • Model name: yolo11n-coco-torch

  • Model source: https://docs.ultralytics.com/models/yolov11/

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 5.35 MB

  • Exposes embeddings? no

  • Tags: detection, coco, torch, yolo, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.3.0

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolo11n-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolo11n-seg-coco-torch#

Edge model producing object outlines directly on phones and edge devices.

Details

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.3.0

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolo11n-seg-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolo11s-coco-torch#

Fast object detector ideal for systems with limited graphics processing power.

Details

  • Model name: yolo11s-coco-torch

  • Model source: https://docs.ultralytics.com/models/yolov11/

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 18.42 MB

  • Exposes embeddings? no

  • Tags: detection, coco, torch, yolo, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.3.0

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolo11s-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolo11s-seg-coco-torch#

Model creating object masks quickly for real-time segmentation applications.

Details

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.3.0

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolo11s-seg-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolo11x-coco-torch#

Object detector prioritizing accuracy over processing speed for critical applications.

Details

  • Model name: yolo11x-coco-torch

  • Model source: https://docs.ultralytics.com/models/yolov11/

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 109.33 MB

  • Exposes embeddings? no

  • Tags: detection, coco, torch, yolo, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.3.0

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolo11x-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolo11x-seg-coco-torch#

Model delivering high-quality object outlines for professional workflows.

Details

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.3.0

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolo11x-seg-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yoloe11l-seg-torch#

Real-time model creating both object outlines and boxes for any described item.

Details

  • Model name: yoloe11l-seg-torch

  • Model source: https://docs.ultralytics.com/models/yoloe

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 67.69 MB

  • Exposes embeddings? no

  • Tags: instances, torch, yolo, zero-shot, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.3.99

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yoloe11l-seg-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
17
18#
19# Make zero-shot predictions with custom classes
20#
21
22model = foz.load_zoo_model(
23    "yoloe11l-seg-torch",
24    classes=["person", "dog", "cat", "bird", "car", "tree", "chair"],
25)
26
27dataset.apply_model(model, label_field="predictions")
28session.refresh()

yoloe11m-seg-torch#

Model producing masks and boxes for objects described in natural language.

Details

  • Model name: yoloe11m-seg-torch

  • Model source: https://docs.ultralytics.com/models/yoloe

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 57.48 MB

  • Exposes embeddings? no

  • Tags: instances, torch, yolo, zero-shot, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.3.99

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yoloe11m-seg-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
17
18#
19# Make zero-shot predictions with custom classes
20#
21
22model = foz.load_zoo_model(
23    "yoloe11m-seg-torch",
24    classes=["person", "dog", "cat", "bird", "car", "tree", "chair"],
25)
26
27dataset.apply_model(model, label_field="predictions")
28session.refresh()

yoloe11s-seg-torch#

Segments specified classes, generating object outlines and boxes for real-time applications.

Details

  • Model name: yoloe11s-seg-torch

  • Model source: https://docs.ultralytics.com/models/yoloe

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 26.52 MB

  • Exposes embeddings? no

  • Tags: instances, torch, yolo, zero-shot, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.3.99

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yoloe11s-seg-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
17
18#
19# Make zero-shot predictions with custom classes
20#
21
22model = foz.load_zoo_model(
23    "yoloe11s-seg-torch",
24    classes=["person", "dog", "cat", "bird", "car", "tree", "chair"],
25)
26
27dataset.apply_model(model, label_field="predictions")
28session.refresh()

yoloev8l-seg-torch#

Model outlining and boxing any object you describe without specific training.

Details

  • Model name: yoloev8l-seg-torch

  • Model source: https://docs.ultralytics.com/models/yoloe

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 102.43 MB

  • Exposes embeddings? no

  • Tags: instances, torch, yolo, zero-shot, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.3.99

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yoloev8l-seg-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
17
18#
19# Make zero-shot predictions with custom classes
20#
21
22model = foz.load_zoo_model(
23    "yoloev8l-seg-torch",
24    classes=["person", "dog", "cat", "bird", "car", "tree", "chair"],
25)
26
27dataset.apply_model(model, label_field="predictions")
28session.refresh()

yoloev8m-seg-torch#

Model creating masks for any object type you name in text.

Details

  • Model name: yoloev8m-seg-torch

  • Model source: https://docs.ultralytics.com/models/yoloe

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 62.75 MB

  • Exposes embeddings? no

  • Tags: instances, torch, yolo, zero-shot, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.3.99

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yoloev8m-seg-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
17
18#
19# Make zero-shot predictions with custom classes
20#
21
22model = foz.load_zoo_model(
23    "yoloev8m-seg-torch",
24    classes=["person", "dog", "cat", "bird", "car", "tree", "chair"],
25)
26
27dataset.apply_model(model, label_field="predictions")
28session.refresh()

yoloev8s-seg-torch#

Compact model producing outlines for objects described in words on edge devices.

Details

  • Model name: yoloev8s-seg-torch

  • Model source: https://docs.ultralytics.com/models/yoloe

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 29.69 MB

  • Exposes embeddings? no

  • Tags: instances, torch, yolo, zero-shot, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.3.99

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yoloev8s-seg-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
17
18#
19# Make zero-shot predictions with custom classes
20#
21
22model = foz.load_zoo_model(
23    "yoloev8s-seg-torch",
24    classes=["person", "dog", "cat", "bird", "car", "tree", "chair"],
25)
26
27dataset.apply_model(model, label_field="predictions")
28session.refresh()

yolov10l-coco-torch#

Object detector with special optimizations for even faster inference on modern hardware.

Details

  • Model name: yolov10l-coco-torch

  • Model source: https://docs.ultralytics.com/models/yolov10/

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 50.00 MB

  • Exposes embeddings? no

  • Tags: detection, coco, torch, yolo, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.2.0

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov10l-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolov10m-coco-torch#

Balanced detector providing good accuracy and speed for general-purpose object detection tasks.

Details

  • Model name: yolov10m-coco-torch

  • Model source: https://docs.ultralytics.com/models/yolov10/

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 32.09 MB

  • Exposes embeddings? no

  • Tags: detection, coco, torch, yolo, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.2.0

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov10m-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolov10n-coco-torch#

Edge-optimized detector for devices with minimal computing resources available.

Details

  • Model name: yolov10n-coco-torch

  • Model source: https://docs.ultralytics.com/models/yolov10/

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 5.59 MB

  • Exposes embeddings? no

  • Tags: detection, coco, torch, yolo, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.2.0

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov10n-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolov10s-coco-torch#

Fast lightweight detector suitable for systems with limited GPU capabilities and memory.

Details

  • Model name: yolov10s-coco-torch

  • Model source: https://docs.ultralytics.com/models/yolov10/

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 15.85 MB

  • Exposes embeddings? no

  • Tags: detection, coco, torch, yolo, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.2.0

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov10s-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolov10x-coco-torch#

High-accuracy detector for demanding object detection applications and research.

Details

  • Model name: yolov10x-coco-torch

  • Model source: https://docs.ultralytics.com/models/yolov10/

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 61.41 MB

  • Exposes embeddings? no

  • Tags: detection, coco, torch, yolo, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.2.0

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov10x-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolov5l-coco-torch#

Real-time detector producing accurate results quickly for demanding vision applications.

Details

  • Model name: yolov5l-coco-torch

  • Model source: https://pytorch.org/hub/ultralytics_yolov5

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 101.96 MB

  • Exposes embeddings? no

  • Tags: detection, coco, torch, yolo, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov5l-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolov5m-coco-torch#

Real-time detector balancing good accuracy with fast processing speeds.

Details

  • Model name: yolov5m-coco-torch

  • Model source: https://pytorch.org/hub/ultralytics_yolov5

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 48.25 MB

  • Exposes embeddings? no

  • Tags: detection, coco, torch, yolo, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov5m-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolov5n-coco-torch#

Lightweight detector for edge devices needing basic object detection capabilities.

Details

  • Model name: yolov5n-coco-torch

  • Model source: https://pytorch.org/hub/ultralytics_yolov5

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 5.31 MB

  • Exposes embeddings? no

  • Tags: detection, coco, torch, yolo, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov5n-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolov5s-coco-torch#

Real-time detector delivering good results with minimal computational requirements.

Details

  • Model name: yolov5s-coco-torch

  • Model source: https://pytorch.org/hub/ultralytics_yolov5

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 17.72 MB

  • Exposes embeddings? no

  • Tags: detection, coco, torch, yolo, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov5s-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolov5x-coco-torch#

High-accuracy detector offering top precision for applications where quality is critical.

Details

  • Model name: yolov5x-coco-torch

  • Model source: https://pytorch.org/hub/ultralytics_yolov5

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 186.09 MB

  • Exposes embeddings? no

  • Tags: detection, coco, torch, yolo, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov5x-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolov8l-coco-torch#

Real-time detector with advanced architecture for improved object finding in complex scenes.

Details

  • Model name: yolov8l-coco-torch

  • Model source: https://docs.ultralytics.com/models/yolov8/

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 83.70 MB

  • Exposes embeddings? no

  • Tags: detection, coco, torch, yolo, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8l-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolov8l-obb-dotav1-torch#

Specialized detector that finds rotated objects in aerial and satellite imagery accurately.

Details

  • Model name: yolov8l-obb-dotav1-torch

  • Model source: https://docs.ultralytics.com/tasks/obb/

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 85.36 MB

  • Exposes embeddings? no

  • Tags: detection, torch, yolo, polylines, obb, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.1.0

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8l-obb-dotav1-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolov8l-oiv7-torch#

General-purpose detector trained on diverse images recognizing over six hundred object categories.

Details

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8l-oiv7-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolov8l-seg-coco-torch#

Creates precise object outlines for detailed image editing and analysis tasks.

Details

  • Model name: yolov8l-seg-coco-torch

  • Model source: https://docs.ultralytics.com/models/yolov8/

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 88.11 MB

  • Exposes embeddings? no

  • Tags: instances, coco, torch, yolo, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8l-seg-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolov8l-world-torch#

Finds and boxes any object you describe using natural language prompts.

Details

  • Model name: yolov8l-world-torch

  • Model source: https://docs.ultralytics.com/models/yolo-world/

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 91.23 MB

  • Exposes embeddings? no

  • Tags: detection, torch, yolo, zero-shot, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.1.0

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8l-world-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
17
18#
19# Make zero-shot predictions with custom classes
20#
21
22model = foz.load_zoo_model(
23    "yolov8l-world-torch",
24    classes=["person", "dog", "cat", "bird", "car", "tree", "chair"],
25)
26
27dataset.apply_model(model, label_field="predictions")
28session.refresh()

yolov8m-coco-torch#

Detector balancing speed and accuracy for everyday object detection needs.

Details

  • Model name: yolov8m-coco-torch

  • Model source: https://docs.ultralytics.com/models/yolov8/

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 49.70 MB

  • Exposes embeddings? no

  • Tags: detection, coco, torch, yolo, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8m-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolov8m-obb-dotav1-torch#

Finds rotated bounding boxes in aerial images for mapping and surveillance applications.

Details

  • Model name: yolov8m-obb-dotav1-torch

  • Model source: https://docs.ultralytics.com/tasks/obb/

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 50.84 MB

  • Exposes embeddings? no

  • Tags: detection, torch, yolo, polylines, obb, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.1.0

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8m-obb-dotav1-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolov8m-oiv7-torch#

Versatile detector recognizing hundreds of different object types across varied image domains.

Details

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8m-oiv7-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolov8m-seg-coco-torch#

Generates object masks with good balance of speed and quality.

Details

  • Model name: yolov8m-seg-coco-torch

  • Model source: https://docs.ultralytics.com/models/yolov8/

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 52.36 MB

  • Exposes embeddings? no

  • Tags: instances, coco, torch, yolo, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8m-seg-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolov8m-world-torch#

Detector understanding text descriptions to find matching objects in images.

Details

  • Model name: yolov8m-world-torch

  • Model source: https://docs.ultralytics.com/models/yolo-world/

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 55.89 MB

  • Exposes embeddings? no

  • Tags: detection, torch, yolo, zero-shot, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.1.0

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8m-world-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
17
18#
19# Make zero-shot predictions with custom classes
20#
21
22model = foz.load_zoo_model(
23    "yolov8m-world-torch",
24    classes=["person", "dog", "cat", "bird", "car", "tree", "chair"],
25)
26
27dataset.apply_model(model, label_field="predictions")
28session.refresh()

yolov8n-coco-torch#

Edge-optimized detector recognizing common objects on resource-limited devices effectively.

Details

  • Model name: yolov8n-coco-torch

  • Model source: https://docs.ultralytics.com/models/yolov8/

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 6.23 MB

  • Exposes embeddings? no

  • Tags: detection, coco, torch, yolo, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8n-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolov8n-obb-dotav1-torch#

Lightweight detector for finding rotated objects in aerial imagery on edge hardware.

Details

  • Model name: yolov8n-obb-dotav1-torch

  • Model source: https://docs.ultralytics.com/tasks/obb/

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 6.24 MB

  • Exposes embeddings? no

  • Tags: detection, torch, yolo, polylines, obb, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.1.0

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8n-obb-dotav1-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolov8n-oiv7-torch#

Edge-friendly detector recognizing hundreds of object categories on resource-limited devices effectively.

Details

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8n-oiv7-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolov8n-seg-coco-torch#

Edge-optimized model producing object outlines on devices with limited resources..

Details

  • Model name: yolov8n-seg-coco-torch

  • Model source: https://docs.ultralytics.com/models/yolov8/

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 6.73 MB

  • Exposes embeddings? no

  • Tags: instances, coco, torch, yolo, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8n-seg-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolov8s-coco-torch#

Detector offering fast performance on mid-range graphics cards and processors.

Details

  • Model name: yolov8s-coco-torch

  • Model source: https://docs.ultralytics.com/models/yolov8/

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 21.53 MB

  • Exposes embeddings? no

  • Tags: detection, coco, torch, yolo, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8s-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolov8s-obb-dotav1-torch#

Efficiently finds rotated objects in aerial photos for mapping and analysis tasks.

Details

  • Model name: yolov8s-obb-dotav1-torch

  • Model source: https://docs.ultralytics.com/tasks/obb/

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 22.17 MB

  • Exposes embeddings? no

  • Tags: detection, torch, yolo, polylines, obb, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.1.0

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8s-obb-dotav1-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolov8s-oiv7-torch#

Compact detector recognizing diverse object types across many different image categories.

Details

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8s-oiv7-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolov8s-seg-coco-torch#

Fast model creating object masks for real-time image segmentation needs.

Details

  • Model name: yolov8s-seg-coco-torch

  • Model source: https://docs.ultralytics.com/models/yolov8/

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 22.79 MB

  • Exposes embeddings? no

  • Tags: instances, coco, torch, yolo, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8s-seg-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolov8s-world-torch#

Lightweight detector finding objects based on text descriptions for edge applications.

Details

  • Model name: yolov8s-world-torch

  • Model source: https://docs.ultralytics.com/models/yolo-world/

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 25.91 MB

  • Exposes embeddings? no

  • Tags: detection, torch, yolo, zero-shot, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.1.0

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8s-world-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
17
18#
19# Make zero-shot predictions with custom classes
20#
21
22model = foz.load_zoo_model(
23    "yolov8s-world-torch",
24    classes=["person", "dog", "cat", "bird", "car", "tree", "chair"],
25)
26
27dataset.apply_model(model, label_field="predictions")
28session.refresh()

yolov8x-coco-torch#

High-accuracy detector for critical applications where precision matters most.

Details

  • Model name: yolov8x-coco-torch

  • Model source: https://docs.ultralytics.com/models/yolov8/

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 130.53 MB

  • Exposes embeddings? no

  • Tags: detection, coco, torch, yolo, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8x-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolov8x-obb-dotav1-torch#

High-precision detector for rotated objects in aerial and satellite imagery analysis.

Details

  • Model name: yolov8x-obb-dotav1-torch

  • Model source: https://docs.ultralytics.com/tasks/obb/

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 133.07 MB

  • Exposes embeddings? no

  • Tags: detection, torch, yolo, polylines, obb, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.1.0

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8x-obb-dotav1-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolov8x-oiv7-torch#

Accurate general detector recognizing over six hundred different object types.

Details

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8x-oiv7-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolov8x-seg-coco-torch#

High-accuracy model generating detailed object outlines for demanding professional applications.

Details

  • Model name: yolov8x-seg-coco-torch

  • Model source: https://docs.ultralytics.com/models/yolov8/

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 137.40 MB

  • Exposes embeddings? no

  • Tags: instances, coco, torch, yolo, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8x-seg-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolov8x-world-torch#

Open-vocabulary detector with high accuracy for text-based object finding.

Details

  • Model name: yolov8x-world-torch

  • Model source: https://docs.ultralytics.com/models/yolo-world/

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 141.11 MB

  • Exposes embeddings? no

  • Tags: detection, torch, yolo, zero-shot, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.1.0

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov8x-world-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)
17
18#
19# Make zero-shot predictions with custom classes
20#
21
22model = foz.load_zoo_model(
23    "yolov8x-world-torch",
24    classes=["person", "dog", "cat", "bird", "car", "tree", "chair"],
25)
26
27dataset.apply_model(model, label_field="predictions")
28session.refresh()

yolov9c-coco-torch#

Detector enhanced with transformer technology for improved object finding capabilities.

Details

  • Model name: yolov9c-coco-torch

  • Model source: https://docs.ultralytics.com/models/yolov9/

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 49.40 MB

  • Exposes embeddings? no

  • Tags: detection, coco, torch, yolo, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.1.0

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov9c-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolov9c-seg-coco-torch#

Compact model producing both masks and boxes with transformer-enhanced accuracy.

Details

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.1.42

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov9c-seg-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolov9e-coco-torch#

Advanced detector with transformer backbone delivering superior accuracy for complex scenes.

Details

  • Model name: yolov9e-coco-torch

  • Model source: https://docs.ultralytics.com/models/yolov9/

  • Model author: Glenn Jocher, et al.

  • Model license: AGPL-3.0

  • Model size: 112.09 MB

  • Exposes embeddings? no

  • Tags: detection, coco, torch, yolo, official

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.1.0

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov9e-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

yolov9e-seg-coco-torch#

Advanced model creating precise object outlines using enhanced transformer architecture.

Details

Requirements

  • Packages: torch>=1.7.0, torchvision>=0.8.1, ultralytics>=8.1.42

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolov9e-seg-coco-torch")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

zero-shot-classification-transformer-torch#

Finds any object you name in images without requiring training on those specific items.

Details

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12classes = ["person", "dog", "cat", "bird", "car", "tree", "chair"]
13
14model = foz.load_zoo_model(
15    "zero-shot-classification-transformer-torch",
16    classes=classes,
17)
18
19dataset.apply_model(model, label_field="predictions")
20
21session = fo.launch_app(dataset)
22
23# some models make require additional arguments
24# check the Hugging Face docs to see if any are needed
25
26# for example, AltCLIP requires `padding=True` in its processor
27model = foz.load_zoo_model(
28    "zero-shot-classification-transformer-torch",
29    classes=classes,
30    name_or_path="BAAI/AltCLIP",
31    transformers_processor_kwargs={
32        "padding": True,
33    }
34)
35
36dataset.apply_model(model, label_field="predictions")
37
38session = fo.launch_app(dataset)

zero-shot-detection-transformer-torch#

Hugging Face Transformers model for zero-shot object detection.

Details

Requirements

  • Packages: torch, torchvision, transformers

  • CPU support

    • yes

  • GPU support

    • yes

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12classes = ["person", "dog", "cat", "bird", "car", "tree", "chair"]
13
14model = foz.load_zoo_model(
15    "zero-shot-detection-transformer-torch",
16    classes=classes,
17)
18
19dataset.apply_model(model, label_field="predictions")
20
21session = fo.launch_app(dataset)

TensorFlow models#

centernet-hg104-1024-coco-tf2#

CenterNet model from Objects as Points with the Hourglass-104 backbone trained on COCO resized to 1024x1024.

Details

  • Model name: centernet-hg104-1024-coco-tf2

  • Model source: tensorflow/models

  • Model author: Xingyi Zhou, et al.

  • Model license: Apache 2.0

  • Model size: 1.33 GB

  • Exposes embeddings? no

  • Tags: detection, coco, tf2, centernet

Requirements

  • CPU support

    • yes

    • Packages: tensorflow>=2|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu>=2|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("centernet-hg104-1024-coco-tf2")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

centernet-hg104-512-coco-tf2#

CenterNet model from Objects as Points with the Hourglass-104 backbone trained on COCO resized to 512x512.

Details

  • Model name: centernet-hg104-512-coco-tf2

  • Model source: tensorflow/models

  • Model author: Xingyi Zhou, et al.

  • Model license: Apache 2.0

  • Model size: 1.49 GB

  • Exposes embeddings? no

  • Tags: detection, coco, tf2, centernet

Requirements

  • CPU support

    • yes

    • Packages: tensorflow>=2|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu>=2|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("centernet-hg104-512-coco-tf2")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

centernet-mobilenet-v2-fpn-512-coco-tf2#

CenterNet model from Objects as Points with the MobileNetV2 backbone trained on COCO resized to 512x512.

Details

  • Model name: centernet-mobilenet-v2-fpn-512-coco-tf2

  • Model source: tensorflow/models

  • Model author: Xingyi Zhou, et al.

  • Model license: Apache 2.0

  • Model size: 41.98 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf2, centernet, mobilenet

Requirements

  • CPU support

    • yes

    • Packages: tensorflow>=2|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu>=2|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("centernet-mobilenet-v2-fpn-512-coco-tf2")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

centernet-resnet101-v1-fpn-512-coco-tf2#

CenterNet model from Objects as Points with the ResNet-101v1 backbone + FPN trained on COCO resized to 512x512.

Details

  • Model name: centernet-resnet101-v1-fpn-512-coco-tf2

  • Model source: tensorflow/models

  • Model author: Xingyi Zhou, et al.

  • Model license: Apache 2.0

  • Model size: 329.96 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf2, centernet, resnet

Requirements

  • CPU support

    • yes

    • Packages: tensorflow>=2|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu>=2|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("centernet-resnet101-v1-fpn-512-coco-tf2")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

centernet-resnet50-v1-fpn-512-coco-tf2#

CenterNet model from Objects as Points with the ResNet-50-v1 backbone + FPN trained on COCO resized to 512x512.

Details

  • Model name: centernet-resnet50-v1-fpn-512-coco-tf2

  • Model source: tensorflow/models

  • Model author: Xingyi Zhou, et al.

  • Model license: Apache 2.0

  • Model size: 194.61 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf2, centernet, resnet

Requirements

  • CPU support

    • yes

    • Packages: tensorflow>=2|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu>=2|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("centernet-resnet50-v1-fpn-512-coco-tf2")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

centernet-resnet50-v2-512-coco-tf2#

CenterNet model from Objects as Points with the ResNet-50v2 backbone trained on COCO resized to 512x512.

Details

  • Model name: centernet-resnet50-v2-512-coco-tf2

  • Model source: tensorflow/models

  • Model author: Xingyi Zhou, et al.

  • Model license: Apache 2.0

  • Model size: 226.95 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf2, centernet, resnet

Requirements

  • CPU support

    • yes

    • Packages: tensorflow>=2|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu>=2|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("centernet-resnet50-v2-512-coco-tf2")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

deeplabv3-cityscapes-tf#

DeepLabv3+ semantic segmentation model from Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation with Xception backbone trained on the Cityscapes dataset.

Details

  • Model name: deeplabv3-cityscapes-tf

  • Model source: tensorflow/models

  • Model author: Liang-Chieh Chen, et al.

  • Model license: Apache 2.0

  • Model size: 158.04 MB

  • Exposes embeddings? no

  • Tags: segmentation, cityscapes, tf, deeplabv3, legacy

Requirements

  • CPU support

    • yes

    • Packages: tensorflow|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("deeplabv3-cityscapes-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

deeplabv3-mnv2-cityscapes-tf#

DeepLabv3+ semantic segmentation model from Encoder-Decoder with Atrous Separable Convolution for Semantic Image Segmentation with MobileNetV2 backbone trained on the Cityscapes dataset.

Details

  • Model name: deeplabv3-mnv2-cityscapes-tf

  • Model source: tensorflow/models

  • Model author: Liang-Chieh Chen, et al.

  • Model license: Apache 2.0

  • Model size: 8.37 MB

  • Exposes embeddings? no

  • Tags: segmentation, cityscapes, tf, deeplabv3, legacy

Requirements

  • CPU support

    • yes

    • Packages: tensorflow|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("deeplabv3-mnv2-cityscapes-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

efficientdet-d0-512-coco-tf2#

EfficientDet-D0 model from EfficientDet: Scalable and Efficient Object Detection trained on COCO resized to 512x512.

Details

  • Model name: efficientdet-d0-512-coco-tf2

  • Model source: tensorflow/models

  • Model author: Mingxing Tan, et al.

  • Model license: Apache 2.0

  • Model size: 29.31 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf2, efficientdet

Requirements

  • CPU support

    • yes

    • Packages: tensorflow>=2|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu>=2|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("efficientdet-d0-512-coco-tf2")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

efficientdet-d0-coco-tf1#

EfficientDet-D0 model from EfficientDet: Scalable and Efficient Object Detection trained on COCO.

Details

  • Model name: efficientdet-d0-coco-tf1

  • Model source: voxel51/automl

  • Model author: Mingxing Tan, et al.

  • Model license: Apache 2.0

  • Model size: 38.20 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf1, efficientdet, legacy

Requirements

  • CPU support

    • yes

    • Packages: tensorflow>=1.14,<2

  • GPU support

    • yes

    • Packages: tensorflow-gpu>=1.14,<2

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("efficientdet-d0-coco-tf1")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

efficientdet-d1-640-coco-tf2#

EfficientDet-D1 model from EfficientDet: Scalable and Efficient Object Detection trained on COCO resized to 640x640.

Details

  • Model name: efficientdet-d1-640-coco-tf2

  • Model source: tensorflow/models

  • Model author: Mingxing Tan, et al.

  • Model license: Apache 2.0

  • Model size: 49.44 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf2, efficientdet

Requirements

  • CPU support

    • yes

    • Packages: tensorflow>=2|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu>=2|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("efficientdet-d1-640-coco-tf2")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

efficientdet-d1-coco-tf1#

EfficientDet-D1 model from EfficientDet: Scalable and Efficient Object Detection trained on COCO.

Details

  • Model name: efficientdet-d1-coco-tf1

  • Model source: voxel51/automl

  • Model author: Mingxing Tan, et al.

  • Model license: Apache 2.0

  • Model size: 61.64 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf1, efficientdet, legacy

Requirements

  • CPU support

    • yes

    • Packages: tensorflow>=1.14,<2

  • GPU support

    • yes

    • Packages: tensorflow-gpu>=1.14,<2

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("efficientdet-d1-coco-tf1")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

efficientdet-d2-768-coco-tf2#

EfficientDet-D2 model from EfficientDet: Scalable and Efficient Object Detection trained on COCO resized to 768x768.

Details

  • Model name: efficientdet-d2-768-coco-tf2

  • Model source: tensorflow/models

  • Model author: Mingxing Tan, et al.

  • Model license: Apache 2.0

  • Model size: 60.01 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf2, efficientdet

Requirements

  • CPU support

    • yes

    • Packages: tensorflow>=2|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu>=2|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("efficientdet-d2-768-coco-tf2")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

efficientdet-d2-coco-tf1#

EfficientDet-D2 model from EfficientDet: Scalable and Efficient Object Detection trained on COCO.

Details

  • Model name: efficientdet-d2-coco-tf1

  • Model source: voxel51/automl

  • Model author: Mingxing Tan, et al.

  • Model license: Apache 2.0

  • Model size: 74.00 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf1, efficientdet, legacy

Requirements

  • CPU support

    • yes

    • Packages: tensorflow>=1.14,<2

  • GPU support

    • yes

    • Packages: tensorflow-gpu>=1.14,<2

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("efficientdet-d2-coco-tf1")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

efficientdet-d3-896-coco-tf2#

EfficientDet-D3 model from EfficientDet: Scalable and Efficient Object Detection trained on COCO resized to 896x896.

Details

  • Model name: efficientdet-d3-896-coco-tf2

  • Model source: tensorflow/models

  • Model author: Mingxing Tan, et al.

  • Model license: Apache 2.0

  • Model size: 88.56 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf2, efficientdet

Requirements

  • CPU support

    • yes

    • Packages: tensorflow>=2|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu>=2|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("efficientdet-d3-896-coco-tf2")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

efficientdet-d3-coco-tf1#

EfficientDet-D3 model from EfficientDet: Scalable and Efficient Object Detection trained on COCO.

Details

  • Model name: efficientdet-d3-coco-tf1

  • Model source: voxel51/automl

  • Model author: Mingxing Tan, et al.

  • Model license: Apache 2.0

  • Model size: 106.44 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf1, efficientdet, legacy

Requirements

  • CPU support

    • yes

    • Packages: tensorflow>=1.14,<2

  • GPU support

    • yes

    • Packages: tensorflow-gpu>=1.14,<2

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("efficientdet-d3-coco-tf1")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

efficientdet-d4-1024-coco-tf2#

EfficientDet-D4 model from EfficientDet: Scalable and Efficient Object Detection trained on COCO resized to 1024x1024.

Details

  • Model name: efficientdet-d4-1024-coco-tf2

  • Model source: tensorflow/models

  • Model author: Mingxing Tan, et al.

  • Model license: Apache 2.0

  • Model size: 151.15 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf2, efficientdet

Requirements

  • CPU support

    • yes

    • Packages: tensorflow>=2|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu>=2|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("efficientdet-d4-1024-coco-tf2")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

efficientdet-d4-coco-tf1#

EfficientDet-D4 model from EfficientDet: Scalable and Efficient Object Detection trained on COCO.

Details

  • Model name: efficientdet-d4-coco-tf1

  • Model source: voxel51/automl

  • Model author: Mingxing Tan, et al.

  • Model license: Apache 2.0

  • Model size: 175.33 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf1, efficientdet, legacy

Requirements

  • CPU support

    • yes

    • Packages: tensorflow>=1.14,<2

  • GPU support

    • yes

    • Packages: tensorflow-gpu>=1.14,<2

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("efficientdet-d4-coco-tf1")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

efficientdet-d5-1280-coco-tf2#

EfficientDet-D5 model from EfficientDet: Scalable and Efficient Object Detection trained on COCO resized to 1280x1280.

Details

  • Model name: efficientdet-d5-1280-coco-tf2

  • Model source: tensorflow/models

  • Model author: Mingxing Tan, et al.

  • Model license: Apache 2.0

  • Model size: 244.41 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf2, efficientdet

Requirements

  • CPU support

    • yes

    • Packages: tensorflow>=2|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu>=2|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("efficientdet-d5-1280-coco-tf2")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

efficientdet-d5-coco-tf1#

EfficientDet-D5 model from EfficientDet: Scalable and Efficient Object Detection trained on COCO.

Details

  • Model name: efficientdet-d5-coco-tf1

  • Model source: voxel51/automl

  • Model author: Mingxing Tan, et al.

  • Model license: Apache 2.0

  • Model size: 275.81 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf1, efficientdet, legacy

Requirements

  • CPU support

    • yes

    • Packages: tensorflow>=1.14,<2

  • GPU support

    • yes

    • Packages: tensorflow-gpu>=1.14,<2

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("efficientdet-d5-coco-tf1")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

efficientdet-d6-1280-coco-tf2#

EfficientDet-D6 model from EfficientDet: Scalable and Efficient Object Detection trained on COCO resized to 1280x1280.

Details

  • Model name: efficientdet-d6-1280-coco-tf2

  • Model source: tensorflow/models

  • Model author: Mingxing Tan, et al.

  • Model license: Apache 2.0

  • Model size: 375.63 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf2, efficientdet

Requirements

  • CPU support

    • yes

    • Packages: tensorflow>=2|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu>=2|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("efficientdet-d6-1280-coco-tf2")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

efficientdet-d6-coco-tf1#

EfficientDet-D6 model from EfficientDet: Scalable and Efficient Object Detection trained on COCO.

Details

  • Model name: efficientdet-d6-coco-tf1

  • Model source: voxel51/automl

  • Model author: Mingxing Tan, et al.

  • Model license: Apache 2.0

  • Model size: 416.43 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf1, efficientdet, legacy

Requirements

  • CPU support

    • yes

    • Packages: tensorflow>=1.14,<2

  • GPU support

    • yes

    • Packages: tensorflow-gpu>=1.14,<2

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("efficientdet-d6-coco-tf1")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

efficientdet-d7-1536-coco-tf2#

EfficientDet-D7 model from EfficientDet: Scalable and Efficient Object Detection trained on COCO resized to 1536x1536.

Details

  • Model name: efficientdet-d7-1536-coco-tf2

  • Model source: tensorflow/models

  • Model author: Mingxing Tan, et al.

  • Model license: Apache 2.0

  • Model size: 376.20 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf2, efficientdet

Requirements

  • CPU support

    • yes

    • Packages: tensorflow>=2|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu>=2|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("efficientdet-d7-1536-coco-tf2")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

faster-rcnn-inception-resnet-atrous-v2-coco-tf#

Faster R-CNN model from Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks atrous version with Inception backbone trained on COCO.

Details

  • Model name: faster-rcnn-inception-resnet-atrous-v2-coco-tf

  • Model source: tensorflow/models

  • Model author: Shaoqing Ren, et al.

  • Model license: Apache 2.0

  • Model size: 234.46 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf, faster-rcnn, inception, resnet

Requirements

  • CPU support

    • yes

    • Packages: tensorflow|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("faster-rcnn-inception-resnet-atrous-v2-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

faster-rcnn-inception-resnet-atrous-v2-lowproposals-coco-tf#

Faster R-CNN model from Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks atrous version with low-proposals and Inception backbone trained on COCO.

Details

  • Model name: faster-rcnn-inception-resnet-atrous-v2-lowproposals-coco-tf

  • Model source: tensorflow/models

  • Model author: Shaoqing Ren, et al.

  • Model license: Apache 2.0

  • Model size: 234.46 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf, faster-rcnn, inception, resnet, legacy

Requirements

  • CPU support

    • yes

    • Packages: tensorflow|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("faster-rcnn-inception-resnet-atrous-v2-lowproposals-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

faster-rcnn-inception-v2-coco-tf#

Faster R-CNN model from Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks with Inception v2 backbone trained on COCO.

Details

  • Model name: faster-rcnn-inception-v2-coco-tf

  • Model source: tensorflow/models

  • Model author: Shaoqing Ren, et al.

  • Model license: Apache 2.0

  • Model size: 52.97 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf, faster-rcnn, inception

Requirements

  • CPU support

    • yes

    • Packages: tensorflow|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("faster-rcnn-inception-v2-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

faster-rcnn-nas-coco-tf#

Faster R-CNN model from Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks with NAS-net backbone trained on COCO.

Details

  • Model name: faster-rcnn-nas-coco-tf

  • Model source: tensorflow/models

  • Model author: Shaoqing Ren, et al.

  • Model license: Apache 2.0

  • Model size: 404.95 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf, faster-rcnn

Requirements

  • CPU support

    • yes

    • Packages: tensorflow|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("faster-rcnn-nas-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

faster-rcnn-nas-lowproposals-coco-tf#

Faster R-CNN model from Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks with low-proposals and NAS-net backbone trained on COCO.

Details

  • Model name: faster-rcnn-nas-lowproposals-coco-tf

  • Model source: tensorflow/models

  • Model author: Shaoqing Ren, et al.

  • Model license: Apache 2.0

  • Model size: 404.88 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf, faster-rcnn, legacy

Requirements

  • CPU support

    • yes

    • Packages: tensorflow|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("faster-rcnn-nas-lowproposals-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

faster-rcnn-resnet101-coco-tf#

Faster R-CNN model from Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks with ResNet-101 backbone trained on COCO.

Details

  • Model name: faster-rcnn-resnet101-coco-tf

  • Model source: tensorflow/models

  • Model author: Shaoqing Ren, et al.

  • Model license: Apache 2.0

  • Model size: 186.41 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf, faster-rcnn, resnet, legacy

Requirements

  • CPU support

    • yes

    • Packages: tensorflow|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("faster-rcnn-resnet101-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

faster-rcnn-resnet101-lowproposals-coco-tf#

Faster R-CNN model from Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks with low-proposals and ResNet-101 backbone trained on COCO.

Details

  • Model name: faster-rcnn-resnet101-lowproposals-coco-tf

  • Model source: tensorflow/models

  • Model author: Shaoqing Ren, et al.

  • Model license: Apache 2.0

  • Model size: 186.41 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf, faster-rcnn, resnet, legacy

Requirements

  • CPU support

    • yes

    • Packages: tensorflow|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("faster-rcnn-resnet101-lowproposals-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

faster-rcnn-resnet50-coco-tf#

Faster R-CNN model from Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks with ResNet-50 backbone trained on COCO.

Details

  • Model name: faster-rcnn-resnet50-coco-tf

  • Model source: tensorflow/models

  • Model author: Shaoqing Ren, et al.

  • Model license: Apache 2.0

  • Model size: 113.57 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf, faster-rcnn, resnet

Requirements

  • CPU support

    • yes

    • Packages: tensorflow|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("faster-rcnn-resnet50-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

faster-rcnn-resnet50-lowproposals-coco-tf#

Faster R-CNN model from Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks with low-proposals and ResNet-50 backbone trained on COCO.

Details

  • Model name: faster-rcnn-resnet50-lowproposals-coco-tf

  • Model source: tensorflow/models

  • Model author: Shaoqing Ren, et al.

  • Model license: Apache 2.0

  • Model size: 113.57 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf, faster-rcnn, resnet, legacy

Requirements

  • CPU support

    • yes

    • Packages: tensorflow|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("faster-rcnn-resnet50-lowproposals-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

inception-resnet-v2-imagenet-tf1#

Inception v2 model from Rethinking the Inception Architecture for Computer Vision trained on ImageNet.

Details

  • Model name: inception-resnet-v2-imagenet-tf1

  • Model source: tensorflow/models

  • Model author: Christian Szegedy, et al.

  • Model license: Apache 2.0

  • Model size: 213.81 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, tf1, inception, resnet

Requirements

  • CPU support

    • yes

    • Packages: tensorflow<2

  • GPU support

    • yes

    • Packages: tensorflow-gpu<2

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("inception-resnet-v2-imagenet-tf1")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

inception-v4-imagenet-tf1#

Inception v4 model from Inception-v4, Inception-ResNet and the Impact of Residual Connections on Learning trained on ImageNet.

Details

  • Model name: inception-v4-imagenet-tf1

  • Model source: tensorflow/models

  • Model author: Christian Szegedy, et al.

  • Model license: Apache 2.0

  • Model size: 163.31 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, tf1, inception, legacy

Requirements

  • CPU support

    • yes

    • Packages: tensorflow<2

  • GPU support

    • yes

    • Packages: tensorflow-gpu<2

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("inception-v4-imagenet-tf1")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

mask-rcnn-inception-resnet-v2-atrous-coco-tf#

Mask R-CNN model from Mask R-CNN atrous version with Inception backbone trained on COCO.

Details

  • Model name: mask-rcnn-inception-resnet-v2-atrous-coco-tf

  • Model source: tensorflow/models

  • Model author: Kaiming He, et al.

  • Model license: Apache 2.0

  • Model size: 254.51 MB

  • Exposes embeddings? no

  • Tags: instances, coco, tf, mask-rcnn, inception, resnet

Requirements

  • CPU support

    • yes

    • Packages: tensorflow|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("mask-rcnn-inception-resnet-v2-atrous-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

mask-rcnn-inception-v2-coco-tf#

Mask R-CNN model from Mask R-CNN with Inception backbone trained on COCO.

Details

  • Model name: mask-rcnn-inception-v2-coco-tf

  • Model source: tensorflow/models

  • Model author: Kaiming He, et al.

  • Model license: Apache 2.0

  • Model size: 64.03 MB

  • Exposes embeddings? no

  • Tags: instances, coco, tf, mask-rcnn, inception

Requirements

  • CPU support

    • yes

    • Packages: tensorflow|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("mask-rcnn-inception-v2-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

mask-rcnn-resnet101-atrous-coco-tf#

Mask R-CNN model from Mask R-CNN atrous version with ResNet-101 backbone trained on COCO.

Details

  • Model name: mask-rcnn-resnet101-atrous-coco-tf

  • Model source: tensorflow/models

  • Model author: Kaiming He, et al.

  • Model license: Apache 2.0

  • Model size: 211.56 MB

  • Exposes embeddings? no

  • Tags: instances, coco, tf, mask-rcnn, resnet, legacy

Requirements

  • CPU support

    • yes

    • Packages: tensorflow|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("mask-rcnn-resnet101-atrous-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

mask-rcnn-resnet50-atrous-coco-tf#

Mask R-CNN model from Mask R-CNN atrous version with ResNet-50 backbone trained on COCO.

Details

  • Model name: mask-rcnn-resnet50-atrous-coco-tf

  • Model source: tensorflow/models

  • Model author: Kaiming He, et al.

  • Model license: Apache 2.0

  • Model size: 138.29 MB

  • Exposes embeddings? no

  • Tags: instances, coco, tf, mask-rcnn, resnet, legacy

Requirements

  • CPU support

    • yes

    • Packages: tensorflow|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("mask-rcnn-resnet50-atrous-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

mobilenet-v2-imagenet-tf1#

MobileNetV2 model from MobileNetV2: Inverted Residuals and Linear Bottlenecks trained on ImageNet.

Details

  • Model name: mobilenet-v2-imagenet-tf1

  • Model source: tensorflow/models

  • Model author: Mark Sandler, et al.

  • Model license: Apache 2.0

  • Model size: 13.64 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, tf1, mobilenet

Requirements

  • CPU support

    • yes

    • Packages: tensorflow<2

  • GPU support

    • yes

    • Packages: tensorflow-gpu<2

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("mobilenet-v2-imagenet-tf1")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

resnet-v1-50-imagenet-tf1#

ResNet-50 v1 model from Deep Residual Learning for Image Recognition trained on ImageNet.

Details

  • Model name: resnet-v1-50-imagenet-tf1

  • Model source: tensorflow/models

  • Model author: Kaiming He, et al.

  • Model license: Apache 2.0

  • Model size: 97.84 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, tf1, resnet, legacy

Requirements

  • CPU support

    • yes

    • Packages: tensorflow<2

  • GPU support

    • yes

    • Packages: tensorflow-gpu<2

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("resnet-v1-50-imagenet-tf1")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

resnet-v2-50-imagenet-tf1#

ResNet-50 v2 model from Deep Residual Learning for Image Recognition trained on ImageNet.

Details

  • Model name: resnet-v2-50-imagenet-tf1

  • Model source: tensorflow/models

  • Model author: Kaiming He, et al.

  • Model license: Apache 2.0

  • Model size: 97.86 MB

  • Exposes embeddings? yes

  • Tags: classification, embeddings, logits, imagenet, tf1, resnet, legacy

Requirements

  • CPU support

    • yes

    • Packages: tensorflow<2

  • GPU support

    • yes

    • Packages: tensorflow-gpu<2

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("resnet-v2-50-imagenet-tf1")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

rfcn-resnet101-coco-tf#

R-FCN object detection model from R-FCN: Object Detection via Region-based Fully Convolutional Networks with ResNet-101 backbone trained on COCO.

Details

  • Model name: rfcn-resnet101-coco-tf

  • Model source: tensorflow/models

  • Model author: Jifeng Dai, et al.

  • Model license: Apache 2.0

  • Model size: 208.16 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf, rfcn, resnet, legacy

Requirements

  • CPU support

    • yes

    • Packages: tensorflow|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("rfcn-resnet101-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

ssd-inception-v2-coco-tf#

Inception Single Shot Detector model from SSD: Single Shot MultiBox Detector trained on COCO.

Details

  • Model name: ssd-inception-v2-coco-tf

  • Model source: tensorflow/models

  • Model author: Wei Liu, et al.

  • Model license: Apache 2.0

  • Model size: 97.50 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf, ssd, inception

Requirements

  • CPU support

    • yes

    • Packages: tensorflow|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("ssd-inception-v2-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

ssd-mobilenet-v1-coco-tf#

Single Shot Detector model from SSD: Single Shot MultiBox Detector with MobileNetV1 backbone trained on COCO.

Details

  • Model name: ssd-mobilenet-v1-coco-tf

  • Model source: tensorflow/models

  • Model author: Wei Liu, et al.

  • Model license: Apache 2.0

  • Model size: 27.83 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf, ssd, mobilenet, legacy

Requirements

  • CPU support

    • yes

    • Packages: tensorflow|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("ssd-mobilenet-v1-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

ssd-mobilenet-v1-fpn-640-coco17#

MobileNetV1 model from MobileNetV2: Inverted Residuals and Linear Bottlenecks resized to 640x640.

Details

  • Model name: ssd-mobilenet-v1-fpn-640-coco17

  • Model source: tensorflow/models

  • Model author: Mark Sandler, et al.

  • Model license: Apache 2.0

  • Model size: 43.91 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf2, ssd, mobilenet

Requirements

  • CPU support

    • yes

    • Packages: tensorflow>=2|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu>=2|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("ssd-mobilenet-v1-fpn-640-coco17")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

ssd-mobilenet-v1-fpn-coco-tf#

FPN Single Shot Detector model from SSD: Single Shot MultiBox Detector with MobileNetV1 backbone trained on COCO.

Details

  • Model name: ssd-mobilenet-v1-fpn-coco-tf

  • Model source: tensorflow/models

  • Model author: Wei Liu, et al.

  • Model license: Apache 2.0

  • Model size: 48.97 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf, ssd, mobilenet, legacy

Requirements

  • CPU support

    • yes

    • Packages: tensorflow|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("ssd-mobilenet-v1-fpn-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

ssd-mobilenet-v2-320-coco17#

MobileNetV2 model from MobileNetV2: Inverted Residuals and Linear Bottlenecks resized to 320x320.

Details

  • Model name: ssd-mobilenet-v2-320-coco17

  • Model source: tensorflow/models

  • Model author: Mark Sandler, et al.

  • Model license: Apache 2.0

  • Model size: 43.91 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf2, ssd, mobilenet

Requirements

  • CPU support

    • yes

    • Packages: tensorflow>=2|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu>=2|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("ssd-mobilenet-v2-320-coco17")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

ssd-resnet50-fpn-coco-tf#

FPN Single Shot Detector model from SSD: Single Shot MultiBox Detector with ResNet-50 backbone trained on COCO.

Details

  • Model name: ssd-resnet50-fpn-coco-tf

  • Model source: tensorflow/models

  • Model author: Wei Liu, et al.

  • Model license: Apache 2.0

  • Model size: 128.07 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf, ssd, resnet, legacy

Requirements

  • CPU support

    • yes

    • Packages: tensorflow|tensorflow-macos

  • GPU support

    • yes

    • Packages: tensorflow-gpu|tensorflow>=2|tensorflow-macos

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("ssd-resnet50-fpn-coco-tf")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)

vgg16-imagenet-tf1#

VGG-16 model from Very Deep Convolutional Networks for Large-Scale Image Recognition trained on ImageNet.

Details

Requirements

  • CPU support

    • yes

    • Packages: tensorflow<2

  • GPU support

    • yes

    • Packages: tensorflow-gpu<2

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "imagenet-sample",
 6    dataset_name=fo.get_default_dataset_name(),
 7    max_samples=50,
 8    shuffle=True,
 9)
10
11model = foz.load_zoo_model("vgg16-imagenet-tf1")
12
13dataset.apply_model(model, label_field="predictions")
14
15session = fo.launch_app(dataset)

yolo-v2-coco-tf1#

YOLOv2 model from YOLO9000: Better, Faster, Stronger trained on COCO.

Details

  • Model name: yolo-v2-coco-tf1

  • Model source: thtrieu/darkflow

  • Model author: Joseph Redmon, et al.

  • Model license: GPL-3.0

  • Model size: 194.49 MB

  • Exposes embeddings? no

  • Tags: detection, coco, tf1, yolo, legacy

Requirements

  • CPU support

    • yes

    • Packages: tensorflow<2

  • GPU support

    • yes

    • Packages: tensorflow-gpu<2

Example usage

 1import fiftyone as fo
 2import fiftyone.zoo as foz
 3
 4dataset = foz.load_zoo_dataset(
 5    "coco-2017",
 6    split="validation",
 7    dataset_name=fo.get_default_dataset_name(),
 8    max_samples=50,
 9    shuffle=True,
10)
11
12model = foz.load_zoo_model("yolo-v2-coco-tf1")
13
14dataset.apply_model(model, label_field="predictions")
15
16session = fo.launch_app(dataset)