Feat: Remote API Parameters Correction (#2802)

* fix: change to gpt4 turbo * add: params * fix: change to gpt 3.5 turbo * delete: redundant * fix: correct description * version bump * add: params * fix: version bump * delete: deprecated * add: params * add: new model * chore: version bump * fix: version correct * add: params * fix: version bump * fix: change to gpt4 turbo * add: params * fix: change to gpt 3.5 turbo * delete: redundant * fix: correct description * version bump * add: params * fix: version bump * delete: deprecated * add: params * add: new model * chore: version bump * fix: version correct * add: params * fix: version bump * fix: llama2 no longer supported * fix: reverse mistral api * fix: add params * fix: mistral api redundant params * fix: typo * fix: typo * fix: correct context length * fix: remove stop --------- Co-authored-by: Van Pham <64197333+Van-QA@users.noreply.github.com>
janhq · May 4, 2024 · 092a572 · 092a572
1 parent c21bc08
commit 092a572
Show file tree

Hide file tree

Showing 6 changed files with 82 additions and 123 deletions.
diff --git a/extensions/inference-groq-extension/package.json b/extensions/inference-groq-extension/package.json
@@ -1,7 +1,7 @@
 {
   "name": "@janhq/inference-groq-extension",
   "productName": "Groq Inference Engine",
-  "version": "1.0.0",
+  "version": "1.0.1",
   "description": "This extension enables fast Groq chat completion API calls",
   "main": "dist/index.js",
   "module": "dist/module.js",

diff --git a/extensions/inference-groq-extension/resources/models.json b/extensions/inference-groq-extension/resources/models.json
@@ -8,18 +8,18 @@
     "id": "llama3-70b-8192",
     "object": "model",
     "name": "Groq Llama 3 70b",
-    "version": "1.0",
+    "version": "1.1",
     "description": "Groq Llama 3 70b with supercharged speed!",
     "format": "api",
-    "settings": {
-      "text_model": false
-    },
+    "settings": {},
     "parameters": {
       "max_tokens": 8192,
       "temperature": 0.7,
-      "top_p": 1,
-      "stop": null,
-      "stream": true
+      "top_p": 0.95,
+      "stream": true,
+      "stop": [],
+      "frequency_penalty": 0,
+      "presence_penalty": 0
     },
     "metadata": {
       "author": "Meta",
@@ -36,18 +36,18 @@
     "id": "llama3-8b-8192",
     "object": "model",
     "name": "Groq Llama 3 8b",
-    "version": "1.0",
+    "version": "1.1",
     "description": "Groq Llama 3 8b with supercharged speed!",
     "format": "api",
-    "settings": {
-      "text_model": false
-    },
+    "settings": {},
     "parameters": {
       "max_tokens": 8192,
       "temperature": 0.7,
-      "top_p": 1,
-      "stop": null,
-      "stream": true
+      "top_p": 0.95,
+      "stream": true,
+      "stop": [],
+      "frequency_penalty": 0,
+      "presence_penalty": 0
     },
     "metadata": {
       "author": "Meta",
@@ -64,53 +64,25 @@
     "id": "gemma-7b-it",
     "object": "model",
     "name": "Groq Gemma 7b Instruct",
-    "version": "1.0",
+    "version": "1.1",
     "description": "Groq Gemma 7b Instruct with supercharged speed!",
     "format": "api",
-    "settings": {
-      "text_model": false
-    },
+    "settings": {},
     "parameters": {
-      "max_tokens": 4096,
+      "max_tokens": 8192,
       "temperature": 0.7,
-      "top_p": 1,
-      "stop": null,
-      "stream": true
+      "top_p": 0.95,
+      "stream": true,
+      "stop": [],
+      "frequency_penalty": 0,
+      "presence_penalty": 0
     },
     "metadata": {
       "author": "Google",
       "tags": ["General"]
     },
     "engine": "groq"
   },
-  {
-    "sources": [
-      {
-        "url": "https://groq.com"
-      }
-    ],
-    "id": "llama2-70b-4096",
-    "object": "model",
-    "name": "Groq Llama 2 70b",
-    "version": "1.0",
-    "description": "Groq Llama 2 70b with supercharged speed!",
-    "format": "api",
-    "settings": {
-      "text_model": false
-    },
-    "parameters": {
-      "max_tokens": 4096,
-      "temperature": 0.7,
-      "top_p": 1,
-      "stop": null,
-      "stream": true
-    },
-    "metadata": {
-      "author": "Meta",
-      "tags": ["General", "Big Context Length"]
-    },
-    "engine": "groq"
-  },
   {
     "sources": [
       {
@@ -120,18 +92,18 @@
     "id": "mixtral-8x7b-32768",
     "object": "model",
     "name": "Groq Mixtral 8x7b Instruct",
-    "version": "1.0",
+    "version": "1.1",
     "description": "Groq Mixtral 8x7b Instruct is Mixtral with supercharged speed!",
     "format": "api",
-    "settings": {
-      "text_model": false
-    },
+    "settings": {},
     "parameters": {
-      "max_tokens": 4096,
+      "max_tokens": 32768,
       "temperature": 0.7,
-      "top_p": 1,
-      "stop": null,
-      "stream": true
+      "top_p": 0.95,
+      "stream": true,
+      "stop": [],
+      "frequency_penalty": 0,
+      "presence_penalty": 0
     },
     "metadata": {
       "author": "Mistral",

diff --git a/extensions/inference-mistral-extension/package.json b/extensions/inference-mistral-extension/package.json
@@ -1,7 +1,7 @@
 {
   "name": "@janhq/inference-mistral-extension",
   "productName": "MistralAI Inference Engine",
-  "version": "1.0.0",
+  "version": "1.0.1",
   "description": "This extension enables Mistral chat completion API calls",
   "main": "dist/index.js",
   "module": "dist/module.js",

diff --git a/extensions/inference-mistral-extension/resources/models.json b/extensions/inference-mistral-extension/resources/models.json
@@ -8,20 +8,20 @@
     "id": "mistral-small-latest",
     "object": "model",
     "name": "Mistral Small",
-    "version": "1.0",
-    "description": "Mistral Small is the ideal choice for simpe tasks that one can do in builk - like Classification, Customer Support, or Text Generation. It offers excellent performance at an affordable price point.",
+    "version": "1.1",
+    "description": "Mistral Small is the ideal choice for simple tasks (Classification, Customer Support, or Text Generation) at an affordable price.",
     "format": "api",
     "settings": {},
     "parameters": {
-      "max_tokens": 4096,
-      "temperature": 0.7
+      "max_tokens": 32000,
+      "temperature": 0.7,
+      "top_p": 0.95,
+      "stream": true
     },
     "metadata": {
       "author": "Mistral",
       "tags": [
-        "Classification",
-        "Customer Support",
-        "Text Generation"
+        "General"
       ]
     },
     "engine": "mistral"
@@ -32,24 +32,23 @@
         "url": "https://docs.mistral.ai/api/"
       }
     ],
-    "id": "mistral-medium-latest",
+    "id": "mistral-large-latest",
     "object": "model",
-    "name": "Mistral Medium",
-    "version": "1.0",
-    "description": "Mistral Medium is the ideal for intermediate tasks that require moderate reasoning - like Data extraction, Summarizing a Document, Writing a Job Description, or Writing Product Descriptions. Mistral Medium strikes a balance between performance and capability, making it suitable for a wide range of tasks that only require language transformaion",
+    "name": "Mistral Large",
+    "version": "1.1",
+    "description": "Mistral Large is ideal for complex tasks (Synthetic Text Generation, Code Generation, RAG, or Agents).",
     "format": "api",
     "settings": {},
     "parameters": {
-      "max_tokens": 4096,
-      "temperature": 0.7
+      "max_tokens": 32000,
+      "temperature": 0.7,
+      "top_p": 0.95,
+      "stream": true
     },
     "metadata": {
       "author": "Mistral",
       "tags": [
-        "Data extraction",
-        "Summarizing a Document",
-        "Writing a Job Description",
-        "Writing Product Descriptions"
+        "General"
       ]
     },
     "engine": "mistral"
@@ -60,24 +59,23 @@
         "url": "https://docs.mistral.ai/api/"
       }
     ],
-    "id": "mistral-large-latest",
+    "id": "open-mixtral-8x22b",
     "object": "model",
-    "name": "Mistral Large",
-    "version": "1.0",
-    "description": "Mistral Large is ideal for complex tasks that require large reasoning capabilities or are highly specialized - like Synthetic Text Generation, Code Generation, RAG, or Agents.",
+    "name": "Mixtral 8x22B",
+    "version": "1.1",
+    "description": "Mixtral 8x22B is a high-performance, cost-effective model designed for complex tasks.",
     "format": "api",
     "settings": {},
     "parameters": {
-      "max_tokens": 4096,
-      "temperature": 0.7
+      "max_tokens": 32000,
+      "temperature": 0.7,
+      "top_p": 0.95,
+      "stream": true
     },
     "metadata": {
       "author": "Mistral",
       "tags": [
-        "Text Generation",
-        "Code Generation",
-        "RAG",
-        "Agents"
+        "General"
       ]
     },
     "engine": "mistral"

diff --git a/extensions/inference-openai-extension/package.json b/extensions/inference-openai-extension/package.json
@@ -1,7 +1,7 @@
 {
   "name": "@janhq/inference-openai-extension",
   "productName": "OpenAI Inference Engine",
-  "version": "1.0.0",
+  "version": "1.0.1",
   "description": "This extension enables OpenAI chat completion API calls",
   "main": "dist/index.js",
   "module": "dist/module.js",

diff --git a/extensions/inference-openai-extension/resources/models.json b/extensions/inference-openai-extension/resources/models.json
@@ -5,20 +5,25 @@
         "url": "https://openai.com"
       }
     ],
-    "id": "gpt-4",
+    "id": "gpt-4-turbo",
     "object": "model",
     "name": "OpenAI GPT 4",
-    "version": "1.0",
+    "version": "1.1",
     "description": "OpenAI GPT 4 model is extremely good",
     "format": "api",
     "settings": {},
     "parameters": {
       "max_tokens": 4096,
-      "temperature": 0.7
+      "temperature": 0.7,
+      "top_p": 0.95,
+      "stream": true,
+      "stop": [],
+      "frequency_penalty": 0,
+      "presence_penalty": 0
     },
     "metadata": {
       "author": "OpenAI",
-      "tags": ["General", "Big Context Length"]
+      "tags": ["General"]
     },
     "engine": "openai"
   },
@@ -31,43 +36,22 @@
     "id": "gpt-4-vision-preview",
     "object": "model",
     "name": "OpenAI GPT 4 with Vision (Preview)",
-    "version": "1.0",
-    "description": "OpenAI GPT 4 with Vision model is extremely good in preview",
+    "version": "1.1",
+    "description": "OpenAI GPT-4 Vision model features vision understanding capabilities",
     "format": "api",
     "settings": {
       "vision_model": true,
       "textModel": false
     },
     "parameters": {
       "max_tokens": 4096,
-      "temperature": 0.7
+      "temperature": 0.7,
+      "top_p": 0.95,
+      "stream": true
     },
     "metadata": {
       "author": "OpenAI",
-      "tags": ["General", "Big Context Length", "Vision"]
-    },
-    "engine": "openai"
-  },
-  {
-    "sources": [
-      {
-        "url": "https://openai.com"
-      }
-    ],
-    "id": "gpt-3.5-turbo-16k-0613",
-    "object": "model",
-    "name": "OpenAI GPT 3.5 Turbo 16k 0613",
-    "version": "1.0",
-    "description": "OpenAI GPT 3.5 Turbo 16k 0613 model is extremely good",
-    "format": "api",
-    "settings": {},
-    "parameters": {
-      "max_tokens": 4096,
-      "temperature": 0.7
-    },
-    "metadata": {
-      "author": "OpenAI",
-      "tags": ["General", "Big Context Length"]
+      "tags": ["General", "Vision"]
     },
     "engine": "openai"
   },
@@ -80,17 +64,22 @@
     "id": "gpt-3.5-turbo",
     "object": "model",
     "name": "OpenAI GPT 3.5 Turbo",
-    "version": "1.0",
-    "description": "OpenAI GPT 3.5 Turbo model is extremely good",
+    "version": "1.1",
+    "description": "OpenAI GPT 3.5 Turbo model is extremely fast",
     "format": "api",
     "settings": {},
     "parameters": {
       "max_tokens": 4096,
-      "temperature": 0.7
+      "temperature": 0.7,
+      "top_p": 0.95,
+      "stream": true,
+      "stop": [],
+      "frequency_penalty": 0,
+      "presence_penalty": 0
     },
     "metadata": {
       "author": "OpenAI",
-      "tags": ["General", "Big Context Length"]
+      "tags": ["General"]
     },
     "engine": "openai"
   }