Update Dockerfile

Dockerfile

@@ -17,8 +17,7 @@ RUN apt-get update && \
     wget && \
     rm -rf /var/lib/apt/lists/*
     
-# Install Python dependencies 
-# Added hf_transfer for faster, more reliable downloads
+# Install Python dependencies with transfer acceleration
 RUN pip3 install huggingface-hub[hf_transfer] openai gradio duckduckgo-search
 
 # Build llama.cpp
@@ -27,16 +26,18 @@ RUN git clone https://github.com/ggerganov/llama.cpp && \
     cmake -B build -S . -DLLAMA_BUILD_SERVER=ON -DGGML_BLAS=ON -DGGML_BLAS_VENDOR=OpenBLAS -DCMAKE_BUILD_TYPE=Release && \
     cmake --build build --config Release --target llama-server -j $(nproc)
 
-# Download model using huggingface-cli (more stable than wget)
+# Download from official Qwen GGUF repo
+# We target the Q8_0 file for high precision
 RUN mkdir -p /models && \
     HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli download \
-    unsloth/Qwen3-0.6B-GGUF \
-    Qwen3-0.6B-UD-Q8_K_XL.gguf \
+    Qwen/Qwen3-0.6B-GGUF \
+    --include "*q8_0.gguf" \
     --local-dir /models \
-    --local-dir-use-symlinks False && \
-    mv /models/Qwen3-0.6B-UD-Q8_K_XL.gguf /models/model.q8_k_xl.gguf
+    --local-dir-use-symlinks False
+
+# Rename the file to a standard name for our startup script
+RUN find /models -name "*.gguf" -exec mv {} /models/model.gguf \;
 
-# Copy app and startup script
 COPY app.py /app.py
 COPY start.sh /start.sh
 RUN chmod +x /start.sh