How to enable reasoning parser in vllm (gguf)
#2
by
fullstack
- opened
Is there a way to enable reasoning parsing in vllm for this model?
vllm serve ./reka-flash-3.1-rekaquant-q3_k_s.gguf --max-model-len 40000 --disable-log-requests --enable-reasoning --reasoning-parser
https://github.com/vllm-project/vllm/blob/main/vllm/reasoning/deepseek_r1_reasoning_parser.py#L22
https://docs.vllm.ai/en/stable/features/reasoning_outputs.html#tool-calling