[Bug] When using Ollama as a local LLM API provider, the output is empty.

[zhoukuncheng]

Which method you use?

• qwen:7b

Provide your subtitle file

[Up to 21°C] Unnamed Memory - 02 (CR 1920x1080 AVC AAC MKV) [503CA3AB].zip

Target language

简体中文

Additional notes

Screenshot

Describe the bug

The transaction output is empty, but the output tokens are increasing normally.

To Reproduce

• Start local LLM model (qwen:7b) using command ollama run qwen:7b
• API & Host settings: API Host http://localhost:11434/v1 Model qwen:7b
• Start translating; no output is produced, but the number of output tokens is increasing.

[gnehs]

We currently only support the OpenAI API.

[gnehs]

We added a Compatibility Mode in version 1.5.0. You can try it out.

[zhoukuncheng]

Thank you, it's working now.