Update pipeline tag, add library name and GitHub link, fix example import

This PR improves the model card for the Ariadne model by:
- Correcting the `pipeline_tag` from `reinforcement-learning` to `image-text-to-text`, accurately reflecting the model's multimodal input and text generation capabilities.
- Adding `library_name: transformers` to the metadata, which enables the automated "how to use" widget on the Hugging Face Hub, as the model demonstrates compatibility with the `transformers` library.
- Including a direct link to the GitHub repository (`https://github.com/Minghe-Shen/Ariadne`) in the content for easier access to the associated code and project details.
- Adding `import torch` to the sample usage code snippet, which is necessary for the snippet to execute correctly due to its use of `torch_dtype` and `torch.cuda.is_available()`.

README.md

@@ -1,10 +1,11 @@
 ---
-license: mit
-language:
-- en
 base_model:
 - Qwen/Qwen2.5-VL-7B-Instruct
-pipeline_tag: reinforcement-learning
+language:
+- en
+license: mit
+pipeline_tag: image-text-to-text
+library_name: transformers
 ---
 
 # 🧠 Ariadne
@@ -12,10 +13,12 @@ pipeline_tag: reinforcement-learning
 This is the official model checkpoint for the paper:  
 **[Ariadne: A Controllable Framework for Probing and Extending VLM Reasoning Boundaries](https://arxiv.org/abs/2511.00710)**  
 
+Code: https://github.com/Minghe-Shen/Ariadne
+
 ### 🔬 Example
 
 ```python
-
+import torch # Added for torch.bfloat16 and torch.cuda.is_available()
 from transformers import AutoModelForImageTextToText, AutoProcessor
 
 MODEL_ID = "..." # path
@@ -62,4 +65,5 @@ input_len = inputs["input_ids"].shape[1]
 gen_ids = sequences[0, input_len:]
 resp_text = processor.tokenizer.decode(
     gen_ids, skip_special_tokens=True, clean_up_tokenization_spaces=True
-).strip()
+).strip()
+```