diff --git a/systems/x86_64-linux/halo/models.ini b/systems/x86_64-linux/halo/models.ini index bfc3fc9..b35ddd0 100644 --- a/systems/x86_64-linux/halo/models.ini +++ b/systems/x86_64-linux/halo/models.ini @@ -15,16 +15,16 @@ fit = on c = 131072 [coder] -hf = unsloth/Qwen3.6-27B-MTP-GGUF:UD-Q8_K_XL +hf = unsloth/Qwen3.6-27B-MTP-GGUF:Q6_K spec-type = ngram-simple,draft-mtp spec-draft-n-max = 5 -spec-draft-p-min = 0.74 threads-batch = 16 temp = 0.6 top-p = 0.95 top-k = 20 min-p = 0.0 -parallel = 4 -c = 1048576 +presence-penalty = 1.5 +parallel = 2 +c = 524288 load-on-startup = true chat-template-kwargs = {"preserve_thinking": true}