feat: set chat_format to None and updated phi-2 gguf model

limcheekin · Dec 24, 2023 · 7e055e1 · 7e055e1
1 parent 617b178
commit 7e055e1
Show file tree

Hide file tree

Showing 2 changed files with 3 additions and 2 deletions.
diff --git a/phi-2/Dockerfile b/phi-2/Dockerfile
@@ -15,4 +15,4 @@ RUN pip install -U pip setuptools wheel && \
 
 # Download model
 RUN mkdir model && \
-    curl -L https://huggingface.co/radames/phi-2-quantized/resolve/main/model-v2-q80.gguf -o model/gguf-model.bin
+    curl -L https://huggingface.co/TheBloke/phi-2-GGUF/resolve/main/phi-2.Q6_K.gguf -o model/gguf-model.bin
diff --git a/phi-2/fastapi_app.py b/phi-2/fastapi_app.py
@@ -21,7 +21,8 @@ def fastapi_app():
         Settings(
             n_threads=4,
             model="/model/gguf-model.bin",
-            embedding=False
+            embedding=False,
+            chat_format=None
         )
     )
     return app