update readme

README.md

@@ -13,23 +13,61 @@ short_description: RTMO PyTorch Checkpoint Tester
 
 # RTMO PyTorch Checkpoint Tester
 
-This HuggingFace Space runs the RTMO (Real-Time Multi-Person) 2D pose estimation model from OpenMMLab.
+This HuggingFace Space provides a real-time 2D multi-person pose estimation demo using the RTMO model from OpenMMLab, accelerated with ZeroGPU.
+
+## Features
+
+- **Remote Checkpoint Selection**: Choose from multiple pre-trained variants (COCO, BODY7, CrowdPose, retrainable RTMO-s) via a dropdown.
+- **Custom Checkpoint Upload**: Upload your own `.pth` file; the application auto-detects RTMO-t/s/m/l variants.
+- **Threshold Adjustment**: Fine-tune **Bounding Box Threshold** and **NMS Threshold** sliders to refine detections.
+- **Example Images**: Three license-free images with people are included for quick testing via the **Examples** panel.
+- **ZeroGPU Acceleration**: Utilizes the `@spaces.GPU()` decorator for GPU inference on HuggingFace Spaces.
 
 ## Usage
-1. Upload an image via the Gradio UI.
-2. (Optional) Provide a path or URL to your own RTMO PyTorch `.pth` checkpoint. If left blank, the default pretrained weights will be used.
-3. Click **Submit**. The annotated image with keypoints will be displayed.
+
+1. **Upload Image**: Drag-and-drop or select an image in the **Upload Image** component (or choose from **Examples**).
+2. **Select Remote Checkpoint**: Pick a preloaded variant from the dropdown menu.
+3. **(Optional) Upload Your Own Checkpoint**: Provide a `.pth` file to override the remote selection; the model variant is detected automatically.
+4. **Adjust Thresholds**: Set **Bounding Box Threshold** (`bbox_thr`) and **NMS Threshold** (`nms_thr`) to control confidence and suppression behavior.
+5. **Run Inference**: Click **Run Inference**. The annotated image and the active checkpoint name will appear.
+
+## Remote Checkpoints
+
+The following variants are available out of the box:
+
+- `rtmo-s_8xb32-600e_coco`
+- `rtmo-m_16xb16-600e_coco`
+- `rtmo-l_16xb16-600e_coco`
+- `rtmo-t_8xb32-600e_body7`
+- `rtmo-s_8xb32-600e_body7`
+- `rtmo-m_16xb16-600e_body7`
+- `rtmo-l_16xb16-600e_body7`
+- `rtmo-s_8xb32-700e_crowdpose`
+- `rtmo-m_16xb16-700e_crowdpose`
+- `rtmo-l_16xb16-700e_crowdpose`
+- `rtmo-s_coco_retrainable` (from Hugging Face)
+
+## Implementation Details
+
+- **GPU Decorator**: `@spaces.GPU()` marks the `predict` function for GPU execution under ZeroGPU.
+- **Inference API**: Leverages `MMPoseInferencer` from MMPose with `pose2d`, `pose2d_weights`, and category `[0]` for person detection.
+- **Monkey-Patch**: Applies a regex patch to bypass `mmdet`’s MMCV version assertion for compatibility.
+- **Variant Detection**: Inspects `backbone.stem.conv.conv.weight` channels in the checkpoint to select the correct RTMO variant.
+- **Checkpoint Management**: Remote files are downloaded to `/tmp/{key}.pth` on demand; uploads use the provided local path.
+- **Output**: Saves visualization images to `/tmp/vis` and displays the first result in the UI.
 
 ## Files
-- **app.py**: Gradio application script that loads the RTMO model, runs inference, and displays results.
-- **requirements.txt**: Python dependencies, including the patched MMCV build and MMPose.
 
-## Model
-We use the `rtmo` alias defined in MMPose’s model zoo. To override, upload your own checkpoint.
+- **app.py**: Main Gradio application script.
+- **requirements.txt**: Python dependencies, including MMCV and MMPose.
+- **README.md**: This documentation file.
 
 ## Development
-If you need to update dependencies or change the model, modify `requirements.txt` and `app.py` accordingly.
 
-## Todos
-1. Support video input
-2. Support models in ONNX format via rtmlib
+To update dependencies, edit `requirements.txt`. To extend functionality or add new variants, modify `app.py` accordingly.
+
+## Future Plans
+
+1. Support video input streams.
+2. Enable ONNX model inference via `rtmlib`.
+