Update model card with paper, code, and usage info

README.md

@@ -1,10 +1,68 @@
 ---
+pipeline_tag: object-detection
+library_name: pytorch
+license: apache-2.0
 tags:
 - model_hub_mixin
 - pytorch_model_hub_mixin
 ---
 
-This model has been pushed to the Hub using the [PytorchModelHubMixin](https://huggingface.co/docs/huggingface_hub/package_reference/mixins#huggingface_hub.PyTorchModelHubMixin) integration:
-- Code: [More Information Needed]
-- Paper: [More Information Needed]
-- Docs: [More Information Needed]
+# DEIMv2-Atto
+
+DEIMv2-Atto is a real-time object detection model introduced in the paper [Real-Time Object Detection Meets DINOv3](https://huggingface.co/papers/2509.20787). It is the ultra-lightweight entry in the DEIMv2 series, which leverages DINOv3 features to achieve state-of-the-art performance-cost trade-offs.
+
+- **Project Page:** [https://intellindust-ai-lab.github.io/projects/DEIMv2/](https://intellindust-ai-lab.github.io/projects/DEIMv2/)
+- **Repository:** [https://github.com/Intellindust-AI-Lab/DEIMv2](https://github.com/Intellindust-AI-Lab/DEIMv2)
+
+## Model Description
+Benefiting from the simplicity and effectiveness of Dense O2O and MAL, DEIM has become a mainstream training framework for real-time DETRs. DEIMv2-Atto employs HGNetv2 with depth and width pruning to meet strict resource budgets. Together with a simplified decoder and an upgraded Dense O2O, this unified design enables DEIMv2 to achieve a superior performance-cost trade-off, making it suitable for GPU, edge, and mobile deployment.
+
+| Model | Dataset | AP | #Params | GFLOPs | Latency (ms) |
+| :---: | :---: | :---: | :---: | :---: | :---: |
+| **Atto** | COCO | **23.8** | 0.5M | 0.8 | 1.10 |
+
+## Usage
+To use this model, you need to have the DEIMv2 repository code available to define the architecture. You can then load the model from the Hub as follows:
+
+```python
+import torch.nn as nn
+from huggingface_hub import PyTorchModelHubMixin
+
+# Ensure the DEIMv2 components from the official GitHub repo are in your python path
+from engine.backbone import HGNetv2
+from engine.deim import LiteEncoder
+from engine.deim import DEIMTransformer
+from engine.deim.postprocessor import PostProcessor
+
+class DEIMv2(nn.Module, PyTorchModelHubMixin):
+    def __init__(self, config):
+        super().__init__()
+        self.backbone = HGNetv2(**config["HGNetv2"])
+        self.encoder = LiteEncoder(**config["LiteEncoder"])
+        self.decoder = DEIMTransformer(**config["DEIMTransformer"])
+        self.postprocessor = PostProcessor(**config["PostProcessor"])
+
+    def forward(self, x, orig_target_sizes):
+        x = self.backbone(x)
+        x = self.encoder(x)
+        x = self.decoder(x)
+        x = self.postprocessor(x, orig_target_sizes)
+
+        return x
+
+# Load the model from the hub
+model = DEIMv2.from_pretrained("Intellindust/DEIMv2_HGNetv2_ATTO_COCO")
+model.eval()
+```
+
+## Citation
+If you use `DEIMv2` or its methods in your work, please cite the following:
+
+```bibtex
+@article{huang2025deimv2,
+  title={Real-Time Object Detection Meets DINOv3},
+  author={Huang, Shihua and Hou, Yongjie and Liu, Longfei and Yu, Xuanlong and Shen, Xi},
+  journal={arXiv},
+  year={2025}
+}
+```