v2.20.2

What's Changed

1. fix multimodal embedding error
2. supports more text embedding model parameters

v2.20.0

What's Changed

1. support qwen 3 models.

v2.19.5

What's Changed

1. supports model model parameters

v2.19.4

What's Changed

1. supports model model parameters

v2.19.3

What's Changed

1. supports more agent parameters

v2.19.2

What's Changed

1. [[Fix] sambert pre_task_id not working

v2.18.4: Feat vl ocr (#38)

What's Changed

1. support model: qwen-vl-ocr-2025-02-18

v2.18.3

What's Changed

1. supports qwen-omni in stream mode
2. supports multimodal-embedding-v1
3. fix VL model using local file as input

v2.18.2

What's Changed

• Support reasoning_content for DeepSeek models.

v2.18.0

What's Changed

• add translation recognizer.
• add video synthesis.