v0.13.3

What's new in 0.13.3 (2024-07-26)

These are the changes in inference v0.13.3.

New features

• FEAT: GLM4 support stream tool call by @codingl2k1 in #1876
• FEAT: support csg-wukong-chat-v0.1 by @qinxuye in #1916
• FEAT: [UI]Add configuration for image and audio models. by @yiboyasss in #1922
• FEAT: support mistral-nemo-instruct by @qinxuye in #1936
• FEAT: CosyVoice speech by @codingl2k1 in #1881
• FEAT: add llama-3.1, llama-3.1-instruct by @Weaxs in #1932
• FEAT: support mistral-large-instruct by @qinxuye in #1944
• Feat: support for llama 3.1 for vllm by @Phoenix500526 in #1935
• FEAT: add rembg flexible model to remove background of image by @qinxuye in #1917

Enhancements

• ENH: added MLX for llama-3-instruct, codestral, Yi-1.5-chat, internlm2.5-chat by @qinxuye in #1908
• ENH: add gptq for llama-3-instruct by @Phoenix500526 in #1934

Bug fixes

• BUG: fix inpainting and flexible infer due to inner API change by @qinxuye in #1907

Documentation

• DOC: update new models to readme by @qinxuye in #1946

New Contributors

• @Phoenix500526 made their first contribution in #1934

Full Changelog: v0.13.2...v0.13.3