#### NOTE
This is a **Hugging Face dataset**. For large datasets, ensure `huggingface_hub>=1.1.3` to avoid rate limits. Learn more in the <a href="https://docs.voxel51.com/integrations/huggingface.html#loading-datasets-from-the-hub" target="_blank">Hugging Face integration docs</a>.

<a href="https://huggingface.co/datasets/Voxel51/Coursera_lecture_dataset_train" target="_blank">![Hugging Face](https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-Dataset-yellow)</a>

# Dataset Card for Lecture Training Set for Coursera MOOC - Hands Data Centric Visual AI

This dataset is the **training dataset for the in-class lectures** of the Hands-on Data Centric Visual AI Coursera course.

This is a [FiftyOne](https://github.com/voxel51/fiftyone) dataset with 16638 samples.

## Installation

If you haven’t already, install FiftyOne:

```bash
pip install -U fiftyone
```

## Usage

```python
import fiftyone as fo
import fiftyone.utils.huggingface as fouh

# Load the dataset
# Note: other available arguments include 'max_samples', etc
dataset = fouh.load_from_hub("Voxel51/Coursera_lecture_dataset_train")

# Launch the App
session = fo.launch_app(dataset)
```

## Dataset Details

### Dataset Description

This dataset is a modified subset of the [LVIS dataset](https://www.lvisdataset.org/).

The dataset here only contains detections, some of which have been artificially perturbed and altered to demonstrate data centric AI techniques and methodologies for the course.

This dataset has the following labels:

- ‘jacket’
- ‘coat’
- ‘jean’
- ‘trousers’
- ‘short_pants’
- ‘trash_can’
- ‘bucket’
- ‘flowerpot’
- ‘helmet’
- ‘baseball_cap’
- ‘hat’
- ‘sunglasses’
- ‘goggles’
- ‘doughnut’
- ‘pastry’
- ‘onion’
- ‘tomato’

### Dataset Sources [optional]

<!-- Provide the basic links for the dataset. -->
- **Repository:** https://www.lvisdataset.org/
- **Paper:** https://arxiv.org/abs/1908.03195

## Uses

The labels in this dataset have been perturbed to illustrate data centric AI techniques for the Hands-on Data Centric AI Coursera MOOC.

## Dataset Structure

Each image in the dataset comes with detailed annotations in FiftyOne detection format. A typical annotation looks like this:

```python
<Detection: {
    'id': '66a2f24cce2f9d11d98d39f3',
    'attributes': {},
    'tags': [],
    'label': 'trousers',
    'bounding_box': [
        0.5562343750000001,
        0.4614166666666667,
        0.1974375,
        0.29300000000000004,
    ],
    'mask': None,
    'confidence': None,
    'index': None,
}>
```

## Dataset Creation

### Curation Rationale

The selected labels for this dataset is due to the fact that these objects can be confusing to a model. Thus, making them a great choice for demonstrating data centric AI techniques.

[More Information Needed]

### Source Data

This is a subset of the [LVIS dataset.](https://www.lvisdataset.org/)

## Citation

<!-- If there is a paper or blog post introducing the dataset, the APA and Bibtex information for that should go in this section. -->

**BibTeX:**

```bibtex
@inproceedings{gupta2019lvis,
  title={{LVIS}: A Dataset for Large Vocabulary Instance Segmentation},
  author={Gupta, Agrim and Dollar, Piotr and Girshick, Ross},
  booktitle={Proceedings of the {IEEE} Conference on Computer Vision and Pattern Recognition},
  year={2019}
}
```