updated dockerfile to use download the model.

Dockerfile

@@ -7,22 +7,24 @@ WORKDIR /app
 # Install system dependencies
 RUN apt-get update && apt-get install -y \
     build-essential \
+    git \
     && rm -rf /var/lib/apt/lists/*
 
 # Create a non-root user
 RUN useradd -m -u 1000 appuser
 
 # Create cache directories and set permissions
-RUN mkdir -p /.cache/huggingface && \
+RUN mkdir -p /app/models/sentence_transformer && \
+    mkdir -p /app/models/qwen && \
+    mkdir -p /.cache/huggingface && \
     mkdir -p /.cache/torch && \
-    mkdir -p /.cache/sentence_transformers && \
-    chown -R appuser:appuser /.cache && \
-    chmod -R 777 /.cache
+    mkdir -p /.cache/sentence_transformers
 
-# Set environment variables for cache locations
+# Set environment variables for cache and model locations
 ENV HF_HOME="/.cache/huggingface"
 ENV TORCH_HOME="/.cache/torch"
-ENV SENTENCE_TRANSFORMERS_HOME="/.cache/sentence_transformers"
+ENV SENTENCE_TRANSFORMERS_HOME="/app/models/sentence_transformer"
+ENV TRANSFORMERS_OFFLINE=1
 
 # Install Python dependencies
 RUN pip install --no-cache-dir \
@@ -39,12 +41,39 @@ RUN pip install --no-cache-dir \
     huggingface_hub \
     accelerate>=0.26.0
 
-# Copy the necessary files into the container
-COPY plant_data_chunks_and_embeddings.csv /app/plant_data_chunks_and_embeddings.csv
+# Create a script to download models
+RUN echo 'import os\n\
+from sentence_transformers import SentenceTransformer\n\
+from transformers import AutoTokenizer, AutoModelForCausalLM\n\
+\n\
+# Download and save sentence transformer model\n\
+model = SentenceTransformer("sentence-transformers/all-mpnet-base-v2")\n\
+model.save("/app/models/sentence_transformer")\n\
+\n\
+# Download and save Qwen model and tokenizer\n\
+model_name = "Qwen/Qwen2.5-1.5B-Instruct"\n\
+tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)\n\
+model = AutoModelForCausalLM.from_pretrained(model_name, trust_remote_code=True, torch_dtype="auto")\n\
+\n\
+tokenizer.save_pretrained("/app/models/qwen")\n\
+model.save_pretrained("/app/models/qwen")\n\
+print("Models downloaded successfully!")' > download_models.py
+
+# Download models during build
+RUN python download_models.py
+
+# Modify the main.py to use local models
 COPY main.py /app/main.py
+RUN sed -i 's|"sentence-transformers/all-mpnet-base-v2"|"/app/models/sentence_transformer"|g' main.py && \
+    sed -i 's|"Qwen/Qwen2.5-1.5B-Instruct"|"/app/models/qwen"|g' main.py
 
-# Set proper permissions for the app directory
-RUN chown -R appuser:appuser /app
+# Copy the data file
+COPY plant_data_chunks_and_embeddings.csv /app/plant_data_chunks_and_embeddings.csv
+
+# Set proper permissions
+RUN chown -R appuser:appuser /app && \
+    chown -R appuser:appuser /.cache && \
+    chmod -R 755 /app/models
 
 # Switch to non-root user
 USER appuser
@@ -52,8 +81,13 @@ USER appuser
 # Set the environment variable to point to the app directory
 ENV PYTHONPATH=/app
 
-# Expose the port the app will run on
-EXPOSE 5000
+# Expose the port
+EXPOSE 8000
+
+# Reduce model loading time by setting specific environment variables
+ENV OMP_NUM_THREADS=1
+ENV MKL_NUM_THREADS=1
+ENV TORCH_NUM_THREADS=1
 
-# Command to run the app using Uvicorn
-CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "5000"]
+# Command to run the app
+CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "5000", "--timeout-keep-alive", "300"]