fix: correct eos token of llava models

extensions/inference-nitro-extension/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@janhq/inference-cortex-extension",
   "productName": "Cortex Inference Engine",
-  "version": "1.0.19",
+  "version": "1.0.20",
   "description": "This extension embeds cortex.cpp, a lightweight inference engine written in C++. See https://jan.ai.\nAdditional dependencies could be installed to run without Cuda Toolkit installation.",
   "main": "dist/index.js",
   "node": "dist/node/index.cjs.js",

extensions/inference-nitro-extension/resources/models/llava-13b/model.json

@@ -12,7 +12,7 @@
   "id": "llava-13b",
   "object": "model",
   "name": "LlaVa 13B Q4",
-  "version": "1.1",
+  "version": "1.2",
   "description": "LlaVa can bring vision understanding to Jan",
   "format": "gguf",
   "settings": {
@@ -24,7 +24,8 @@
     "mmproj": "mmproj-model-f16.gguf"
   },
   "parameters": {
-    "max_tokens": 4096
+    "max_tokens": 4096,
+    "stop": ["</s>"]
   },
   "metadata": {
     "author": "liuhaotian",

extensions/inference-nitro-extension/resources/models/llava-7b/model.json

@@ -12,7 +12,7 @@
   "id": "llava-7b",
   "object": "model",
   "name": "LlaVa 7B",
-  "version": "1.1",
+  "version": "1.2",
   "description": "LlaVa can bring vision understanding to Jan",
   "format": "gguf",
   "settings": {
@@ -24,7 +24,8 @@
     "mmproj": "mmproj-model-f16.gguf"
   },
   "parameters": {
-    "max_tokens": 4096
+    "max_tokens": 4096,
+    "stop": ["</s>"]
   },
   "metadata": {
     "author": "liuhaotian",