model: add Gemma 2 models

mlc-ai · Aug 1, 2024 · eca7b95 · eca7b95
1 parent da9dce0
commit eca7b95
Show file tree

Hide file tree

Showing 4 changed files with 51 additions and 16 deletions.
diff --git a/app/constant.ts b/app/constant.ts
@@ -626,13 +626,13 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     },
   },
   {
-    name: "gemma-2b-it-q4f16_1-MLC",
+    name: "gemma-2-2b-it-q4f16_1-MLC",
     display_name: "Gemma",
     provider: "Google",
     size: "2B",
     quantization: "q4f16_1",
     family: "Gemma",
-    vram_required_MB: 1476.52,
+    vram_required_MB: 1895.3,
     low_resource_required: false,
     buffer_size_required_bytes: 262144000,
     required_features: ["shader-f16"],
@@ -644,14 +644,13 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     },
   },
   {
-    name: "gemma-2b-it-q4f32_1-MLC",
+    name: "gemma-2-2b-it-q4f32_1-MLC",
     display_name: "Gemma",
     provider: "Google",
     size: "2B",
     quantization: "q4f32",
-
     family: "Gemma",
-    vram_required_MB: 1750.66,
+    vram_required_MB: 2508.75,
     low_resource_required: false,
     buffer_size_required_bytes: 262144000,
     recommended_config: {
@@ -662,13 +661,13 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     },
   },
   {
-    name: "gemma-2b-it-q4f16_1-MLC-1k",
+    name: "gemma-2-2b-it-q4f16_1-MLC-1k",
     display_name: "Gemma",
     provider: "Google",
     size: "2B",
     quantization: "q4f16_1",
     family: "Gemma",
-    vram_required_MB: 1476.52,
+    vram_required_MB: 1583.3,
     low_resource_required: true,
     buffer_size_required_bytes: 262144000,
     required_features: ["shader-f16"],
@@ -680,13 +679,13 @@ export const DEFAULT_MODELS: ModelRecord[] = [
     },
   },
   {
-    name: "gemma-2b-it-q4f32_1-MLC-1k",
+    name: "gemma-2-2b-it-q4f32_1-MLC-1k",
     display_name: "Gemma",
     provider: "Google",
     size: "2B",
     quantization: "q4f32",
     family: "Gemma",
-    vram_required_MB: 1750.66,
+    vram_required_MB: 1884.75,
     low_resource_required: true,
     buffer_size_required_bytes: 262144000,
     recommended_config: {
@@ -696,6 +695,42 @@ export const DEFAULT_MODELS: ModelRecord[] = [
       top_p: 0.95,
     },
   },
+  {
+    name: "gemma-2-9b-it-q4f16_1-MLC",
+    display_name: "Gemma",
+    provider: "Google",
+    size: "9B",
+    quantization: "q4f16",
+    family: "Gemma",
+    vram_required_MB: 6422.01,
+    low_resource_required: false,
+    buffer_size_required_bytes: 262144000,
+    required_features: ["shader-f16"],
+    recommended_config: {
+      temperature: 0.7,
+      presence_penalty: 0,
+      frequency_penalty: 1,
+      top_p: 0.95,
+    },
+  },
+  {
+    name: "gemma-2-9b-it-q4f32_1-MLC",
+    display_name: "Gemma",
+    provider: "Google",
+    size: "9B",
+    quantization: "q4f32",
+    family: "Gemma",
+    vram_required_MB: 8383.33,
+    low_resource_required: false,
+    buffer_size_required_bytes: 262144000,
+    required_features: ["shader-f16"],
+    recommended_config: {
+      temperature: 0.7,
+      presence_penalty: 0,
+      frequency_penalty: 1,
+      top_p: 0.95,
+    },
+  },
   {
     name: "stablelm-2-zephyr-1_6b-q4f16_1-MLC",
     display_name: "StableLM",

diff --git a/app/store/config.ts b/app/store/config.ts
@@ -208,9 +208,9 @@ export const useAppConfig = createPersistStore(
   }),
   {
     name: StoreKey.Config,
-    version: 0.47,
+    version: 0.48,
     migrate: (persistedState, version) => {
-      if (version < 0.47) {
+      if (version < 0.48) {
         return {
           ...DEFAULT_CONFIG,
           ...(persistedState as any),

diff --git a/package.json b/package.json
@@ -17,7 +17,7 @@
   "dependencies": {
     "@fortaine/fetch-event-source": "^3.0.6",
     "@hello-pangea/dnd": "^16.5.0",
-    "@mlc-ai/web-llm": "^0.2.51",
+    "@mlc-ai/web-llm": "^0.2.52",
     "@serwist/next": "^9.0.2",
     "@svgr/webpack": "^6.5.1",
     "emoji-picker-react": "^4.9.2",

diff --git a/yarn.lock b/yarn.lock
@@ -1180,10 +1180,10 @@
     "@jridgewell/resolve-uri" "^3.1.0"
     "@jridgewell/sourcemap-codec" "^1.4.14"
 
-"@mlc-ai/web-llm@^0.2.51":
-  version "0.2.51"
-  resolved "https://registry.yarnpkg.com/@mlc-ai/web-llm/-/web-llm-0.2.51.tgz#73dced138262bf79a1bc082dd36dc6d1a928b54f"
-  integrity sha512-62aiDyiSzUah1vqnTcZvH3lFRjz4VM2CeIBp7QpcFgX2HPA3vVYhDbV0cOVGAfK1MfH1CipZiyjOjkQUGjjRTw==
+"@mlc-ai/web-llm@^0.2.52":
+  version "0.2.52"
+  resolved "https://registry.yarnpkg.com/@mlc-ai/web-llm/-/web-llm-0.2.52.tgz#15c0496b3ce8c35bd0ac5d1dd92d30cdc6c055db"
+  integrity sha512-TRjAQsOCUf7vrwVeOZoBZtB9KnW9InWK2w85rJWL7IBSSW/znYsNdD3/opmEGtf7NTKr4RBIu+8RdnhcZoHk4A==
   dependencies:
     loglevel "^1.9.1"