LMstudio error
"error rendering prompt with jinja template: "Error: Parser Error: Expected closing statement token. OpenSquareBracket !== CloseStatement."
"Studio\resources\app.webpack\lib\llmworker.js:114:232891)". This is usually an issue with the model's prompt template. If you are using a popular model, you can try to search the model under lmstudio-community, which will have fixed prompt templates. If you cannot find one, you are welcome to post this issue to our discord or issue tracker on GitHub. Alternatively, if you know how to write jinja templates, you can override the prompt template in My Models > model settings > Prompt Template."
The template from QwQ-32b works.
It seems like template from QWQ doesn't contain generated by transformers library additional formatting for enabling/disabling thinking.
The template from QwQ-32b works.
How does it work and how much of an upgrade over the qwq32b
It seems like template from QWQ doesn't contain generated by transformers library additional formatting for enabling/disabling thinking.
Yes, it's just a temporary solution, but it's better than the template of ChatLM. In the test, it can exclude the thinking content in the context.
The template from QwQ-32b works.
How does it work and how much of an upgrade over the qwq32b
It can disable reasoning by setting "/no_thinking" in the system prompt; reasoning is enabled by default. Unfortunately, my GPU only has 24GB of VRAM, so I can only run the Q4 quantized version and thus cannot conduct an accurate evaluation comparison. My current impression is that it is more stable than QwQ, but the reasoning process can still be quite long at times.