Add library_name, pipeline_tag, and links to paper and code

README.md

@@ -1,17 +1,21 @@
 ---
+datasets:
+- keentomato/human_behavior_atlas
 language: en
 license: apache-2.0
 tags:
-  - human-behavior
-  - multimodal
-  - qwen2.5-omni
-datasets:
-  - keentomato/human_behavior_atlas
+- human-behavior
+- multimodal
+- qwen2.5-omni
+pipeline_tag: any-to-any
+library_name: transformers
 ---
 
 # OmniSapiens SFT
 
-Fine-tuned [Qwen2.5-Omni-7B](https://huggingface.co/Qwen/Qwen2.5-Omni-7B) for human behavior understanding.
+Fine-tuned [Qwen2.5-Omni-7B](https://huggingface.co/Qwen/Qwen2.5-Omni-7B) for human behavior understanding. This model was introduced in the paper [OmniSapiens: A Foundation Model for Social Behavior Processing via Heterogeneity-Aware Relative Policy Optimization](https://huggingface.co/papers/2602.10635).
+
+**Official Code:** [https://github.com/MIT-MI/human_behavior_atlas](https://github.com/MIT-MI/human_behavior_atlas)
 
 ## Benchmark
 Evaluated on [keentomato/human_behavior_atlas](https://huggingface.co/datasets/keentomato/human_behavior_atlas).
@@ -99,4 +103,4 @@ with torch.no_grad():
 
 answer = processor.decode(generated[0][inputs["input_ids"].shape[1]:], skip_special_tokens=True)
 print(answer)
-```
+```