defaults to alpaca conv template

Migrate the gradio server to openai v1 (#3032)

Gradio Web Server for Multimodal Models (#2960)

Update gradio demo and API model providers (#3030)

Update gradio version in `pyproject.toml` and fix a bug (#3029)

Update readme (#3028)

Update MLX integration to use new generate_step function signature (#3021)

Upgrade gradio to 4.17 (#3027)

Update lightllm_integration.md (#3014)

Update model_support.md (#3016)

code update (#2997)

Fix type hint for play_a_match_single (#3008)

fix: inconsistent tokenization by llama tokenizer (#3006)

fix: tokenization mismatch when training with different templates (#2996)

update yuan2.0 generation (#2989)