chore(halo): add fast model

2026-05-22 09:10:30 +02:00 · 2026-05-22 09:10:30 +02:00 · fc3e40686a
commit fc3e40686a
parent 6fd6060dd7
1 changed files with 15 additions and 0 deletions
--- a/systems/x86_64-linux/halo/models.ini
+++ b/systems/x86_64-linux/halo/models.ini
@ -41,3 +41,18 @@ parallel            = 2
 c                   = 524288
 load-on-startup     = true
 chat-template-kwargs = {"preserve_thinking": true}
+
+[fast]
+hf                  = byteshape/Qwen3.6-35B-A3B-MTP-GGUF:IQ4_XS
+spec-type           = ngram-simple,draft-mtp
+spec-draft-n-max    = 3
+threads-batch       = 16
+temp                = 0.6
+top-p               = 0.95
+top-k               = 20
+min-p               = 0.0
+presence-penalty    = 1.5
+parallel            = 1
+c                   = 131072
+chat-template-kwargs = {"preserve_thinking": true}
+