edtorre/ qwen3.6-hermes:latest

27 5 days ago

Qwen 3.6 27B (Q4_K_M) optimized for Hermes Agent — 64K context, 8192 max tokens, MTP for speed, flash attention + Q8 KV cache.

vision tools thinking
... /
params
bf746530d62c · 152B
{
"draft_num_predict": 4,
"min_p": 0,
"num_ctx": 65536,
"num_predict": 8192,
"presence_penalty": 1.2,
"repeat_penalty": 1,
"temperature": 0.7,
"top_k": 20,
"top_p": 0.9
}