QwQ-32B-Preview LoRA for separating thinking/answer parts

This LoRA file was fine-tuned to make QwQ constantly separate its private thoughts from the final answer using <THINKING>...</THINKING><ANSWER>...</ANSWER> tags.

A Q4_K_M GGUF version (which can be used as an adapter for Ollama) is available on shakedzy/QwQ-32B-Preview-with-Tags-LoRA-GGUF.

Downloads last month
7
Safetensors
Model size
18.9B params
Tensor type
F32
FP16
U8
Inference API
Unable to determine this model鈥檚 pipeline type. Check the docs .

Model tree for shakedzy/QwQ-32b-Preview-bnb-4bit-wTags

Base model

Qwen/Qwen2.5-32B
Adapter
(24)
this model