Add pipeline tag and license to model card, link to code

This PR adds the missing `pipeline_tag` and `license` to the model card metadata. The `pipeline_tag` is set to `image-text-to-text` based on the model's functionality as described in the abstract and usage example.
It also adds a link to the github repository so it's easier for people to use the model.

README.md

@@ -1,6 +1,8 @@
 ---
 library_name: transformers
 tags: []
+pipeline_tag: image-text-to-text
+license: mit
 ---
 
 # Fine-Grained Visual Classification on HAM10000
@@ -9,16 +11,15 @@ Project Page: [SelfSynthX](https://github.com/sycny/SelfSynthX).
 
 Paper on arXiv: [Enhancing Cognition and Explainability of Multimodal Foundation Models with Self-Synthesized Data](https://arxiv.org/abs/2502.14044)
 
-
-This model is a fine-tuned multimodal foundation model developed on the [LLaVA-1.5-7B-hf](https://huggingface.co/llava-hf/llava-1.5-7b-hf) base, optimized for fine-grained skin lesion classification and explainability using the HAM10000 dataset.
+This model is a fine-tuned multimodal foundation model developed on the [LLaVA-1.5-7B-hf](https://huggingface.co/llava-hf/llava-1.5-7B-hf) base, optimized for fine-grained skin lesion classification and explainability using the HAM10000 dataset.
 
 ## Key Details
 
-- **Base Model:** LLaVA-1.5-7B  
+- **Base Model:** LLaVA-1.5-7B
 - **Dataset:** HAM10000
-- **Innovation:**  
-  - **Self-Synthesized Data:** Generates interpretable explanations by extracting lesion-specific visual concepts using the Information Bottleneck principle.  
-  - **Iterative Fine-Tuning:** Uses reward model-free rejection sampling to progressively improve classification accuracy and explanation quality.  
+- **Innovation:**
+  - **Self-Synthesized Data:** Generates interpretable explanations by extracting lesion-specific visual concepts using the Information Bottleneck principle.
+  - **Iterative Fine-Tuning:** Uses reward model-free rejection sampling to progressively improve classification accuracy and explanation quality.
 - **Intended Use:** Skin lesion classification with human-verifiable explanations for dermatological analysis.
 
 ## How to Use
@@ -31,8 +32,8 @@ from transformers import AutoProcessor, LlavaForConditionalGeneration
 
 model_id = "YuchengShi/LLaVA-v1.5-7B-HAM10000"
 model = LlavaForConditionalGeneration.from_pretrained(
-    model_id, 
-    torch_dtype=torch.float16, 
+    model_id,
+    torch_dtype=torch.float16,
     low_cpu_mem_usage=True,
 ).to("cuda")
 processor = AutoProcessor.from_pretrained(model_id)
@@ -57,7 +58,7 @@ print(processor.decode(output[0][2:], skip_special_tokens=True))
 
 ## Training & Evaluation
 
-- **Training:** Fine-tuned using LoRA on HAM10000 with iterative rejection sampling.  
+- **Training:** Fine-tuned using LoRA on HAM10000 with iterative rejection sampling.
 - **Evaluation:** Demonstrates higher accuracy and robust, interpretable explanations compared to baseline models.
 
 ## Citation
@@ -73,4 +74,10 @@ If you use this model, please cite:
   year={2025},
   url={https://openreview.net/forum?id=lHbLpwbEyt}
 }
-```
+```
+
+## Contact
+
+For any questions, suggestions, or issues, please open an issue on GitHub or contact us at [[email protected]](mailto:[email protected]).
+
+Github repository: https://github.com/sycny/SelfSynthX