Add Q3_K_M variand of Qwen3.5-9B

2026-03-06 23:21:58 +01:00
parent 28cb53c031
commit 42038207fc
1 changed files with 28 additions and 0 deletions
--- a/apps/llama/configs/config.yaml
+++ b/apps/llama/configs/config.yaml
@@ -677,3 +677,31 @@ models:
        --no-warmup
        --port ${PORT}
        --chat-template-kwargs "{\"enable_thinking\": false}"
+
+  "Qwen3.5-9B-GGUF:Q3_K_M":
+    ttl: 600
+    cmd: |
+      /app/llama-server
+        -hf unsloth/Qwen3.5-9B-GGUF:Q3_K_M
+        --ctx-size 16384
+        --temp 0.6
+        --top-p 0.95
+        --top-k 20
+        --min-p 0.00
+        --no-warmup
+        --port ${PORT}
+        --chat-template-kwargs "{\"enable_thinking\": true}"
+
+  "Qwen3.5-9B-GGUF-nothink:Q3_K_M":
+    ttl: 600
+    cmd: |
+      /app/llama-server
+        -hf unsloth/Qwen3.5-9B-GGUF:Q3_K_M
+        --ctx-size 16384
+        --temp 0.6
+        --top-p 0.95
+        --top-k 20
+        --min-p 0.00
+        --no-warmup
+        --port ${PORT}
+        --chat-template-kwargs "{\"enable_thinking\": false}"