oneformer-cityscapes-swin-large-instance-torch

OneFormer with Swin-L backbone trained on Cityscapes for instance segmentation, returning per-instance binary masks.

Details

• Model name: oneformer-cityscapes-swin-large-instance-torch

• Model source: https://huggingface.co/shi-labs/oneformer_cityscapes_swin_large

• Model author: Jitesh Jain, et al.

• Model license: Apache-2.0

• Model size: 1.64 GB

• Exposes embeddings? no

• Tags: instances, torch, transformers, official

Requirements

• Packages: torch, torchvision, transformers

• CPU support

  • yes

• GPU support

  • yes

Example usage

import fiftyone as fo
import fiftyone.zoo as foz

dataset = foz.load_zoo_dataset(
    "coco-2017",
    split="validation",
    dataset_name=fo.get_default_dataset_name(),
    max_samples=50,
    shuffle=True,
)

model = foz.load_zoo_model("oneformer-cityscapes-swin-large-instance-torch")

dataset.apply_model(model, label_field="predictions")

session = fo.launch_app(dataset)