Improve model card: Add metadata, abstract, and GitHub link for AEPO
This PR significantly enhances the model card for the Agentic Entropy-Balanced Policy Optimization (AEPO) model. It adds essential metadata: library_name: transformers to enable direct usage via the Hugging Face transformers library, and pipeline_tag: robotics to correctly classify the model's functionality for agentic learning and tool-use. Additional tags (reinforcement-learning, agents, tool-use, web-agents, qwen) are also included to improve model discoverability.
Furthermore, this update integrates the paper's abstract into the model card content for better understanding of the model, and provides an explicit link to the official GitHub repository for the codebase. In adherence to guidelines, no sample usage code snippet is added as none was found in the project's GitHub README.