Merge pull request #935 from janhq/fix/inference_model_parameters

fix: model parameters for inference extensions

extensions/inference-nitro-extension/src/helpers/sse.ts

@@ -16,7 +16,7 @@ export function requestInference(
       messages: recentMessages,
       model: model.id,
       stream: true,
-      // ...model.parameters,
+      ...model.parameters,
     });
     fetch(INFERENCE_URL, {
       method: "POST",

extensions/inference-openai-extension/src/helpers/sse.ts

@@ -22,8 +22,8 @@ export function requestInference(
     const requestBody = JSON.stringify({
       messages: recentMessages,
       stream: true,
-      model: model_id
-      // ...model.parameters,
+      model: model_id,
+      ...model.parameters,
     });
     fetch(`${engine.full_url}`, {
       method: "POST",

models/gpt-3.5-turbo-16k-0613/model.json

@@ -7,9 +7,7 @@
     "description": "OpenAI GPT 3.5 Turbo 16k 0613 model is extremely good",
     "format": "api",
     "settings": {},
-    "parameters": {
-      "max_tokens": 4096
-    },
+    "parameters": {},
     "metadata": {
       "author": "OpenAI",
       "tags": ["General", "Big Context Length"]

models/gpt-4/model.json

@@ -2,14 +2,12 @@
     "source_url": "https://openai.com",
     "id": "gpt-4",
     "object": "model",
-    "name": "OpenAI GPT 3.5",
+    "name": "OpenAI GPT 4",
     "version": 1.0,
-    "description": "OpenAI GPT 3.5 model is extremely good",
+    "description": "OpenAI GPT 4 model is extremely good",
     "format": "api",
     "settings": {},
-    "parameters": {
-      "max_tokens": 4096
-    },
+    "parameters": {},
     "metadata": {
       "author": "OpenAI",
       "tags": ["General", "Big Context Length"]