apps/librechat/release.yaml

@@ -54,7 +54,10 @@ spec:
                   "gemma3-4b",
                   "gemma3-4b-novision",
                   "Qwen3-4B-Thinking-2507",
-                  "Qwen3-4B-Thinking-2507-long-ctx"
+                  "Qwen3-4B-Thinking-2507-long-ctx",
+                  "Qwen2.5-VL-7B-Instruct-GGUF",
+                  "Qwen2.5-VL-32B-Instruct-GGUF-IQ1_S",
+                  "Qwen2.5-VL-32B-Instruct-GGUF-Q2_K_L"
                 ]
               titleConvo: true
               titleModel: "gemma3-4b-novision"

apps/llama/configs/config.yaml

@@ -175,3 +175,42 @@ models:
       --flash-attn
       --cache-type-k q8_0 --cache-type-v q8_0
       --port ${PORT}
+  "Qwen2.5-VL-32B-Instruct-GGUF-IQ1_S":
+    ttl: 600
+    cmd: |
+      /app/llama-server
+      -hf unsloth/Qwen2.5-VL-32B-Instruct-GGUF:IQ1_S
+      -ngl 99 -c 16384 --predict 8192
+      --temp 0.7
+      --min-p 0.00
+      --top-p 0.8
+      --top-k 20
+      --repeat-penalty 1.0
+      --no-warmup
+      --port ${PORT}
+  "Qwen2.5-VL-32B-Instruct-GGUF-Q2_K_L":
+    ttl: 600
+    cmd: |
+      /app/llama-server
+      -hf unsloth/Qwen2.5-VL-32B-Instruct-GGUF:Q2_K_L
+      -ngl 99 -c 16384 --predict 8192
+      --temp 0.7
+      --min-p 0.00
+      --top-p 0.8
+      --top-k 20
+      --repeat-penalty 1.0
+      --no-warmup
+      --port ${PORT}
+  "Qwen2.5-VL-7B-Instruct-GGUF":
+    ttl: 600
+    cmd: |
+      /app/llama-server
+      -hf unsloth/Qwen2.5-VL-7B-Instruct-GGUF:Q4_K_M
+      -ngl 37 -c 16384 --predict 8192
+      --temp 0.7
+      --min-p 0.00
+      --top-p 0.8
+      --top-k 20
+      --repeat-penalty 1.0
+      --no-warmup
+      --port ${PORT}