v1.7.0

xorbitsai/inferencev1.7.0Jun 13, 2025by XprobeBot

AI Summary

A major feature release adding support for CogView4, MiniCPM4, Qwen3, and DeepSeek-R1 models, along with UI enhancements including Japanese and Korean language support.

Key Highlights

  • Support for CogView4 image generation model
  • Support for MiniCPM4 Series and Qwen3 embedding/reranker
  • UI support for Japanese and Korean languages
  • DeepSeek-R1-0528 Mixed quantization support

New Features

  • CogView4 image model support
  • UI support for model_ability filter for image and video models
  • UI auto-switch to active tab when Running Models page loads
  • Support for first-last-frame in video models
  • Japanese and Korean language support in UI
  • SeACoParaformer model support
  • verbose_json support for FunASR audio2text models
  • DeepSeek-R1-0528 Mixed quantization support
  • Support for engines in embedding models
  • MiniCPM4 Series support
  • model_engine parameter for embedding models in UI
  • kwargs for transcripts client API
  • Qwen3 embedding support
  • Qwen3-reranker support
  • PCM response_format support

Full Release Notes

# What's new in 1.7.0 (2025-06-13)
These are the changes in inference v1.7.0.
## New features
* FEAT: support CogView4 image model by @qinxuye in https://github.com/xorbitsai/inference/pull/3557
* FEAT: [UI] support model_ability filter for image and video models. by @yiboyasss in https://github.com/xorbitsai/inference/pull/3563
* FEAT: [UI] auto-switch to active tab when Running Models page loads. by @yiboyasss in https://github.com/xorbitsai/inference/pull/3568
* FEAT: support first-last-frame to video by @qinxuye in https://github.com/xorbitsai/inference/pull/3555
* FEAT: [UI] add Japanese and Korean language support. by @yiboyasss in https://github.com/xorbitsai/inference/pull/3574
* FEAT: SeACoParaformer model by @leslie2046 in https://github.com/xorbitsai/inference/pull/3587
* FEAT: support verbose_json for funasr family audio2text models by @leslie2046 in https://github.com/xorbitsai/inference/pull/3591
* FEAT: support deepseek-r1-0528 Mixed quantization by @Jun-Howie in https://github.com/xorbitsai/inference/pull/3601
* FEAT: support engines for embedding models by @pengjunfeng11 in https://github.com/xorbitsai/inference/pull/2791
* FEAT:support MiniCPM4 Series by @Jun-Howie in https://github.com/xorbitsai/inference/pull/3609
* FEAT: [UI] add model_engine parameter to embedding model. by @yiboyasss in https://github.com/xorbitsai/inference/pull/3617
* FEAT: add kwargs for transripts client API by @leslie2046 in https://github.com/xorbitsai/inference/pull/3622
* FEAT: support qwen3 embedding by @qinxuye in https://github.com/xorbitsai/inference/pull/3615
* FEAT: support qwen3-reranker by @qinxuye in https://github.com/xorbitsai/inference/pull/3627
## Enhancements
* ENH: Support pcm response_format by @codingl2k1 in https://github.com/xorbitsai/inference/pull/3606
## Bug fixes
* BUG: Fix dependency by @codingl2k1 in https://github.com/xorbitsai/inference/pull/3566
* BUG: Fix cmdline by @codingl2k1 in https://github.com/xorbitsai/inference/pull/3589
* BUG: fix potential hang for sglang by @qinxuye in https://github.com/xorbitsai/inference/pull/3597
* BUG: [UI] fixed the mobile language switching bug. by @yiboyasss in https://github.com/xorbitsai/inference/pull/3608
* BUG:  Fix the error when using Qwen function call with Spring AI. by @aniya105 in https://github.com/xorbitsai/inference/pull/3614
## Documentation
* DOC: update links by @qinxuye in https://github.com/xorbitsai/inference/pull/3565
* DOC: Update CosyVoice doc by @codingl2k1 in https://github.com/xorbitsai/inference/pull/3605
* DOC: update models by @qinxuye in https://github.com/xorbitsai/inference/pull/3628
## Others
* FIX: [UI] fix model_engine parameter bug. by @yiboyasss in https://github.com/xorbitsai/inference/pull/3620

## New Contributors
* @aniya105 made their first contribution in https://github.com/xorbitsai/inference/pull/3614

**Full Changelog**: https://github.com/xorbitsai/inference/compare/v1.6.1...v1.7.0