From 5f78b90618db1ae2d86fe8867fd1493ea15e7d59 Mon Sep 17 00:00:00 2001
From: Nathan Sarrazin <sarrazin.nathan@gmail.com>
Date: Wed, 21 Feb 2024 14:00:02 +0100
Subject: [PATCH] Add gemma model to prod config (#854)

* Add `google/gemma-7b-it` to production config

* Add gemma prompt format to PROMPTS.md

* Make sampling parameters optional
---
 .env.template            | 29 ++++++++++++++++++++++++++++-
 PROMPTS.md               |  6 ++++++
 src/lib/server/models.ts |  4 ++--
 3 files changed, 36 insertions(+), 3 deletions(-)

diff --git a/.env.template b/.env.template
index b069e572..5ea6ce34 100644
--- a/.env.template
+++ b/.env.template
@@ -31,7 +31,34 @@ MODELS=`[
       }
     ]
   },
-    {
+  {
+    "name" : "google/gemma-7b-it",
+    "description": "Gemma 7B is from a family of lightweight models from Google built from the same research and technology used to create the Gemini models.",
+    "websiteUrl" : "https://add-link-to-their-comms-here.com/",
+    "logoUrl": "https://huggingface.co/datasets/huggingchat/models-logo/resolve/main/google-logo.png",
+    "modelUrl": "https://huggingface.co/google/gemma-7b-it",
+    "preprompt": "",
+    "chatPromptTemplate" : "{{#each messages}}{{#ifUser}}<start_of_turn>user\n{{#if @first}}{{#if @root.preprompt}}{{@root.preprompt}}\n{{/if}}{{/if}}{{content}}<end_of_turn>\n<start_of_turn>model\n{{/ifUser}}{{#ifAssistant}}{{content}}<end_of_turn>\n{{/ifAssistant}}{{/each}}",
+    "promptExamples": [
+      {
+        "title": "Write an email from bullet list",
+        "prompt": "As a restaurant owner, write a professional email to the supplier to get these products every week: \n\n- Wine (x10)\n- Eggs (x24)\n- Bread (x12)"
+      }, {
+        "title": "Code a snake game",
+        "prompt": "Code a basic snake game in python, give explanations for each step."
+      }, {
+        "title": "Assist in a task",
+        "prompt": "How do I make a delicious lemon cheesecake?"
+      }
+    ],
+    "parameters": {
+        "do_sample": true,
+        "truncate": 7168,
+        "max_new_tokens": 1024,
+        "stop" : ["<end_of_turn>"]
+      }
+  },
+  {
       "name": "meta-llama/Llama-2-70b-chat-hf",
       "description": "The latest and biggest model from Meta, fine-tuned for chat.",
       "logoUrl": "https://huggingface.co/datasets/huggingchat/models-logo/resolve/main/meta-logo.png",
diff --git a/PROMPTS.md b/PROMPTS.md
index 4f921a6d..1847e0dc 100644
--- a/PROMPTS.md
+++ b/PROMPTS.md
@@ -61,3 +61,9 @@ System: {{preprompt}}\nUser:{{#each messages}}{{#ifUser}}{{content}}\nFalcon:{{/
 ```env
 <s>{{#if @root.preprompt}}Source: system\n\n {{@root.preprompt}} <step> {{/if}}{{#each messages}}{{#ifUser}}Source: user\n\n {{content}} <step> {{/ifUser}}{{#ifAssistant}}Source: assistant\n\n {{content}} <step> {{/ifAssistant}}{{/each}}Source: assistant\nDestination: user\n\n ``
 ```
+
+## Gemma
+
+```env
+{{#each messages}}{{#ifUser}}<start_of_turn>user\n{{#if @first}}{{#if @root.preprompt}}{{@root.preprompt}}\n{{/if}}{{/if}}{{content}}<end_of_turn>\n<start_of_turn>model\n{{/ifUser}}{{#ifAssistant}}{{content}}<end_of_turn>\n{{/ifAssistant}}{{/each}}
+```
diff --git a/src/lib/server/models.ts b/src/lib/server/models.ts
index e280eb4d..588baf0c 100644
--- a/src/lib/server/models.ts
+++ b/src/lib/server/models.ts
@@ -58,9 +58,9 @@ const modelConfig = z.object({
 	endpoints: z.array(endpointSchema).optional(),
 	parameters: z
 		.object({
-			temperature: z.number().min(0).max(1),
+			temperature: z.number().min(0).max(1).optional(),
 			truncate: z.number().int().positive().optional(),
-			max_new_tokens: z.number().int().positive(),
+			max_new_tokens: z.number().int().positive().optional(),
 			stop: z.array(z.string()).optional(),
 			top_p: z.number().positive().optional(),
 			top_k: z.number().positive().optional(),