tweak 8

Dockerfile

@@ -23,7 +23,12 @@ ENV GRADIO_SERVER_NAME="0.0.0.0"
 
 # Download a checkpoint
 RUN mkdir content
-ADD --chown=user  https://huggingface.co/LoneStriker/OpenBioLLM-Llama3-8B-GGUF/blob/main/OpenBioLLM-Llama3-8B-Q5_K_M.gguf content/OpenBioLLM-Llama3-8B-Q5_K_M.gguf
+# Download the model using the Hugging Face CLI
+
+# Install the Hugging Face CLI tool
+RUN pip install huggingface_hub
+
+RUN huggingface-cli download NicholasJohn/OpenBioLLM-Llama3-8B-Q5_K_M.gguf --local-dir ./models
 
 RUN ls -l $HOME/app/content
 

app.py

@@ -11,7 +11,7 @@ from huggingface_hub.file_download import http_get
 llm = Llama(
     # model_path="./Phi-3-mini-4k-instruct-q4.gguf",
     # model_path="./llama3-gguf/OpenBioLLM-Llama3-8B-Q5_K_M.gguf",
-    model_path = "./content/OpenBioLLM-Llama3-8B-Q5_K_M.gguf",
+    model_path = "./models/OpenBioLLM-Llama3-8B-Q5_K_M.gguf",
     n_ctx=2048,
     n_gpu_layers=50, # change n_gpu_layers if you have more or less VRAM 
 )

requirements.txt

@@ -1,3 +1,3 @@
-llama-cpp-python==0.2.69
-huggingface-hub==0.23.0
+# llama-cpp-python==0.2.69
+# huggingface-hub==0.23.0
 gradio==4.28.3