- Added windows llama proxy

- Added terraform-mcp
- Added S3 storage to gitea
- Changed context size for windows LLMs
- Changed number of cores for VMs, added P2000 support

ansible/host_vars/ci-01.lab.alexpires.me.yml

@@ -42,18 +42,38 @@ llama_server_models:
     model_id: "mradermacher/Luau-Qwen3-4B-FIM-v0.1-GGUF"
     quant: "IQ4_XS"
     revision: "main"
+    preset:
+      ctx-size: 4096
   - name: "ministral-3-3b-instruct"
     model_id: "unsloth/Ministral-3-3B-Instruct-2512-GGUF"
     quant: "IQ4_XS"
     revision: "main"
+    preset:
+      ctx-size: 4096
   - name: "qwen2.5-coder-1.5b-instruct"
     model_id: "Qwen/Qwen2.5-Coder-1.5B-Instruct-GGUF"
     quant: "Q4_K_M"
     revision: "main"
-  - name: "qwen2.5-coder-3b-instruct"
-    model_id: "unsloth/Qwen2.5-Coder-3B-Instruct-128K-GGUF"
+    preset:
+      ctx-size: 4096
+  - name: "qwencoder-3b"
+    model_id: "Jojobigworld/XiYanSQL-QwenCoder-3B-2502-Q4_K_M-GGUF"
     quant: "Q4_K_M"
     revision: "main"
+    preset:
+      ctx-size: 4096
+  - name: "qwen2.5-coder-1.5b-q8"
+    model_id: "ggml-org/Qwen2.5-Coder-1.5B-Q8_0-GGUF"
+    quant: "Q8_0"
+    revision: "main"
+    preset:
+      ctx-size: 4096
+  - name: "qwen2.5-coder-3b-q8"
+    model_id: "ggml-org/Qwen2.5-Coder-3B-Q8_0-GGUF"
+    quant: "Q8_0"
+    revision: "main"
+    preset:
+      ctx-size: 4096
 
 llama_server_host: "0.0.0.0"