updated to single server

2026-07-15 10:30:35 +02:00
parent a08333b70a
commit a993838826
4 changed files with 63 additions and 61 deletions
@@ -7,7 +7,7 @@ REPO_DIR := $(shell pwd)
 ITEMS := skills tools agents

 # LLM services to manage
-LLM_SERVICES := atlas forge swift
+LLM_SERVICES := odin

 PLIST_PATH := ~/Library/LaunchAgents

@@ -68,4 +68,4 @@ restart-llm-%:
 	@launchctl bootout gui/$$(id -u) $(PLIST_PATH)/com.vllm-mlx.$*.plist || true
 	@launchctl bootstrap gui/$$(id -u) $(PLIST_PATH)/com.vllm-mlx.$*.plist

-restart-llm: $(patsubst %,restart-llm-%,$(LLM_SERVICES))
+restart-llm: $(patsubst %,restart-llm-%,$(LLM_SERVICES))
@@ -0,0 +1,26 @@
+manager:
+  memory_budget: 450
+  contention_policy:
+    strategy: wait_then_preempt
+    wait_timeout_s: 45
+    preempt_after_s: 15
+
+models:
+  - name: Qwen-Coder-Next
+    model: mlx-community/Qwen3-Coder-Next-4bit
+    estimated_memory_gb: 50
+    reasoning_parser: qwen3
+
+  - name: Qwen3.6
+    model: mlx-community/Qwen3.6-35B-A3B-6bit
+    estimated_memory_gb: 40
+    reasoning_parser: qwen3
+
+  - name: gemma
+    model: mlx-community/gemma-4-31b-it-8bit
+    estimated_memory_gb: 40
+    reasoning_parser: gemma4
+
+  - name: MiniMax-M3
+    model: pipenetwork/MiniMax-M3-MLX-mixed-3_6bit
+    estimated_memory_gb: 200
@@ -1,59 +0,0 @@
-{
-  "$schema": "https://opencode.ai/config.json",
-  "provider": {
-    "swift": {
-      "npm": "@ai-sdk/anthropic",
-      "name": "Odin",
-      "options": {
-        "baseURL": "http://192.168.2.49:12000/v1",
-        "apiKey": "changeme"
-      },
-      "models": {
-        "mlx-community/Qwen3-Coder-Next-4bit": {
-          "name": "Odin Swift",
-          "tool_call": true,
-          "options": {
-            "temperature": 1.0,
-            "top_p": 0.95
-          }
-        }
-      }
-    },
-    "forge": {
-      "npm": "@ai-sdk/anthropic",
-      "name": "Odin",
-      "options": {
-        "baseURL": "http://192.168.2.49:12001/v1",
-        "apiKey": "changeme"
-      },
-      "models": {
-        "mlx-community/Qwen3-Coder-Next-6bit": {
-          "name": "Odin Forge",
-          "tool_call": true,
-          "options": {
-            "temperature": 1.0,
-            "top_p": 0.95
-          }
-        }
-      }
-    },
-    "atlas": {
-      "npm": "@ai-sdk/anthropic",
-      "name": "Odin",
-      "options": {
-        "baseURL": "http://192.168.2.49:12002/v1",
-        "apiKey": "changeme"
-      },
-      "models": {
-        "mlx-community/Qwen3.5-35B-A3B-6bit": {
-          "name": "Odin Atlas",
-          "tool_call": true,
-          "options": {
-            "temperature": 1.0,
-            "top_p": 0.95
-          }
-        }
-      }
-    }
-  }
-}
@@ -0,0 +1,35 @@
+{
+  "$schema": "https://opencode.ai/config.json",
+  "provider": {
+    "odin": {
+      "npm": "@ai-sdk/anthropic",
+      "name": "Odin",
+      "options": {
+        "baseURL": "http://192.168.2.49:12000/v1",
+        "apiKey": "changeme"
+      },
+      "models": {
+        "Qwen-Coder-Next": {
+          "name": "Qwen-Coder-Next",
+          "tool_call": true,
+          "options": {
+            "temperature": 1.0,
+            "top_p": 0.95
+          }
+        },
+        "Qwen3.6": {
+          "name": "Qwen3.6",
+          "tool_call": true
+        },
+        "gemma": {
+          "name": "gemma",
+          "tool_call": true
+        },
+        "MiniMax-M3": {
+          "name": "MiniMax-M3",
+          "tool_call": true
+        }
+      }
+    }
+  }
+}