Chore: model hub v0.5.1 update (#3036)

* init model

* init qwen2

* version bump

* refactor: correct icon

* chore: Refactor/issue template feature request (#3037)

* refactor: add issue template form for bug

* refactor: config blank_issues_enabled: false

* refactor: config feature request

* refactor: config feature request

---------

Co-authored-by: Van-QA <van@jan.ai>

* refactor: correct icon

* refactor: allow blank issue

---------

Co-authored-by: Van-QA <van@jan.ai>
Co-authored-by: Van Pham <64197333+Van-QA@users.noreply.github.com>

.github/ISSUE_TEMPLATE/config.yml

@@ -1,5 +1,5 @@
 ## To encourage contributors to use issue templates, we don't allow blank issues
-blank_issues_enabled: false
+blank_issues_enabled: true
 
 contact_links:
   - name: "\u2753 Our GitHub Discussions page"

.github/ISSUE_TEMPLATE/documentation-request.md

@@ -1,5 +1,5 @@
 ---
-name: "\U+1F56E Documentation request"
+name: "📖 Documentation request"
 about: Documentation requests
 title: 'docs: TITLE'
 labels: 'type: documentation'

.github/ISSUE_TEMPLATE/epic-request.md

@@ -1,7 +1,7 @@
 ---
-name: Epic request
+name: "💥 Epic request"
 about: Suggest an idea for this project
-title: '\U0001F4A5 epic: [DESCRIPTION]'
+title: 'epic: [DESCRIPTION]'
 labels: 'type: epic'
 assignees: ''
 

extensions/inference-nitro-extension/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@janhq/inference-cortex-extension",
   "productName": "Cortex Inference Engine",
-  "version": "1.0.11",
+  "version": "1.0.12",
   "description": "This extension embeds cortex.cpp, a lightweight inference engine written in C++. See https://nitro.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/qwen2-7b/model.json

@@ -0,0 +1,36 @@
+{
+    "sources": [
+      {
+        "filename": "Qwen2-7B-Instruct-Q4_K_M.gguf",
+        "url": "https://huggingface.co/bartowski/Qwen2-7B-Instruct-GGUF/resolve/main/Qwen2-7B-Instruct-Q4_K_M.gguf"
+      }
+    ],
+    "id": "qwen2-7b",
+    "object": "model",
+    "name": "Qwen 2 Instruct 7B Q4",
+    "version": "1.0",
+    "description": "Qwen is optimized at Chinese, ideal for everyday tasks.",
+    "format": "gguf",
+    "settings": {
+      "ctx_len": 32768,
+      "prompt_template": "<|im_start|>system\n{system_message}<|im_end|>\n<|im_start|>user\n{prompt}<|im_end|>\n<|im_start|>assistant",
+      "llama_model_path": "Qwen2-7B-Instruct-Q4_K_M.gguf",
+      "ngl": 28
+    },
+    "parameters": {
+      "temperature": 0.7,
+      "top_p": 0.95,
+      "stream": true,
+      "max_tokens": 32768,
+      "stop": [],
+      "frequency_penalty": 0,
+      "presence_penalty": 0
+    },
+    "metadata": {
+      "author": "Alibaba",
+      "tags": ["7B", "Finetuned"],
+      "size": 4680000000
+    },
+    "engine": "nitro"
+  }
+  

extensions/inference-nitro-extension/rollup.config.ts

@@ -39,6 +39,8 @@ const aya8bJson = require('./resources/models/aya-23-8b/model.json')
 const aya35bJson = require('./resources/models/aya-23-35b/model.json')
 const phimediumJson = require('./resources/models/phi3-medium/model.json')
 const codestralJson = require('./resources/models/codestral-22b/model.json')
+const qwen2Json = require('./resources/models/qwen2-7b/model.json')
+
 
 export default [
   {
@@ -84,7 +86,8 @@ export default [
           phimediumJson,
           aya8bJson,
           aya35bJson,
-          codestralJson
+          codestralJson,
+          qwen2Json
         ]),
         NODE: JSON.stringify(`${packageJson.name}/${packageJson.node}`),
         DEFAULT_SETTINGS: JSON.stringify(defaultSettingJson),