Update model card with IJCV paper information

README.md

@@ -4,7 +4,10 @@ language:
 - en
 tags:
 - robotics
-- grasping
+- robotic-grasping
+- target-driven-grasping
+- occlusion
+- 6dof-grasping
 - 3d-vision
 - shape-completion
 - pytorch
@@ -13,18 +16,27 @@ library_name: pytorch
 
 # TARGO-Net
 
-TARGO-Net is a PyTorch checkpoint release for target-oriented robotic grasping in cluttered scenes. This repository currently provides the model weights needed by the cleaned TARGO inference and training code.
+This repository hosts the released checkpoints for **TARGO-Net**, the model from:
 
-## Files
+**TARGO and TARGO-Net: Benchmarking Target-Driven Object Grasping Under Occlusions**  
+Accepted at **International Journal of Computer Vision (IJCV), 2026**.
+
+- Project page: https://targo-benchmark.github.io/
+- Paper DOI: https://doi.org/10.1007/s11263-025-02716-9
+- arXiv: https://arxiv.org/abs/2407.06168
+
+## Overview
+
+TARGO studies target-driven 6-DoF robotic grasping under visual occlusion. Given a single RGB-D observation and a target object, TARGO-Net uses target shape completion and target-scene feature fusion to predict collision-aware grasp poses that remain robust when the target is partially occluded by clutter.
+
+## Checkpoints
 
 | File | Description |
 | --- | --- |
-| `checkpoints/targonet.pt` | TARGO grasp prediction network checkpoint. |
-| `checkpoints/adapointr.pth` | AdaPoinTr shape completion checkpoint used by the TARGO inference pipeline. |
+| `checkpoints/targonet.pt` | TARGO-Net grasp prediction checkpoint. |
+| `checkpoints/adapointr.pth` | AdaPoinTr target shape completion checkpoint used by the TARGO-Net inference pipeline. |
 
-## Usage
-
-Download the checkpoints with `huggingface_hub`:
+## Download
 
 ```python
 from huggingface_hub import snapshot_download
@@ -36,20 +48,20 @@ snapshot_download(
 )
 ```
 
-For the cleaned TARGO code, place or link the downloaded files as:
+For the cleaned TARGO code, place or link the files as:
 
 ```text
 checkpoints/targonet.pt
 checkpoints/adapointr.pth
 ```
 
-Then run inference with the processed VGN-format test data:
+## Inference
 
 ```bash
 python inference_targo.py   --model checkpoints/targonet.pt   --sc_model_path checkpoints/adapointr.pth   --test_root /path/to/processed_vgn/test_set_gaussian_0.002
 ```
 
-The `--test_root` directory should contain:
+The processed VGN-format `--test_root` should contain:
 
 ```text
 scenes/
@@ -57,14 +69,24 @@ mesh_pose_dict/
 occ_level_dict.json
 ```
 
-## Training / Evaluation Data
+## Data
+
+The benchmark/dataset files are not included in this model repository. Please see the project page for code, data, and benchmark details:
 
-The checkpoints are intended for the TARGO/VGN-style data pipeline. The processed dataset is not included in this model repository.
+https://targo-benchmark.github.io/
 
-## Limitations
+## Citation
 
-These weights are research artifacts and are intended for offline robotics research workflows. Real-robot deployment requires additional calibration, safety checks, and environment-specific validation.
+```bibtex
+@article{xia2026targo,
+  title={TARGO and TARGO-Net: Benchmarking Target-Driven Object Grasping Under Occlusions},
+  author={Xia, Yan and Ding, Ran and Qin, Ziyuan and Zhan, Guanqi and Zhou, Kaichen and Yang, Long and Dong, Hao and Cremers, Daniel},
+  journal={International Journal of Computer Vision},
+  year={2026},
+  doi={10.1007/s11263-025-02716-9}
+}
+```
 
 ## License
 
-MIT.
+The model repository is released under the MIT license. Please also check the licenses of the benchmark data and any third-party assets used in your experiments.