Fix code llama

kaarthik108 · Sep 25, 2023 · 8e1d610 · 8e1d610
1 parent 0030c34
commit 8e1d610
Show file tree

Hide file tree

Showing 2 changed files with 7 additions and 4 deletions.
diff --git a/.gitignore b/.gitignore
@@ -3,4 +3,7 @@ __pycache__/
 secrets.toml
 .ruff_cache
 
-archived_logs/
+archived_logs/
+
+build/
+snowchat.egg-info/
diff --git a/chain.py b/chain.py
@@ -57,7 +57,7 @@
 context: \n {context}
 <</CONTEXT>>
 
-WRITE RESPONSES IN MARKDOWN FORMAT and code in ```sql  ```
+write responses in markdown format
 
 Answer:
 
@@ -74,8 +74,8 @@
 supabase_key = st.secrets["SUPABASE_SERVICE_KEY"]
 supabase: Client = create_client(supabase_url, supabase_key)
 
-VERSION = "da5676342de1a5a335b848383af297f592b816b950a43d251a0a9edd0113604b"
-LLAMA = "replicate/codellama-13b-instruct:{}".format(VERSION)
+VERSION = "be553392065353425e0f0193d2a896d6a5ff201549f5d7cd9180c8dfdeac39ed"
+LLAMA = "meta/codellama-13b-instruct:{}".format(VERSION)
 
 
 def get_chain_replicate(vectorstore, callback_handler=None):