Update Dockerfile

Dockerfile

@@ -2,35 +2,26 @@ FROM python:3.9-slim
 
 WORKDIR /code
 
-# System deps for audio (minimize with --no-install-recommends)
-RUN apt-get update && apt-get install -y --no-install-recommends ffmpeg libsndfile1 && rm -rf /var/lib/apt/lists/*
-
-# Install CPU-only Torch first (to satisfy optimum dep without CUDA bloat)
-RUN pip install --no-cache-dir torch --index-url https://download.pytorch.org/whl/cpu
-
-# Copy requirements
+# Copy requirements (make sure it includes soundfile)
 COPY ./requirements.txt /code/requirements.txt
 
-# Then install the rest (Torch dep is now satisfied)
-RUN pip install --no-cache-dir -r /code/requirements.txt
+# Install system deps for ffmpeg + audio libs
+RUN apt-get update && apt-get install -y ffmpeg libsndfile1 && rm -rf /var/lib/apt/lists/*
 
-# Hugging Face cache inside container (clean + writable)
+# Install Python deps
+RUN pip install --no-cache-dir --upgrade -r /code/requirements.txt
+
+# Use HF_HOME to control transformers cache location
 ENV HF_HOME=/code/hf_cache
 RUN mkdir -p /code/hf_cache && chmod -R 777 /code/hf_cache
 
-# Pre-download ONNX model + processor at build time
+# Pre-download ASR model at build time (so runtime doesn't download/cache)
 RUN python - <<'PY'
-from transformers import WhisperProcessor
-from optimum.onnxruntime import ORTModelForSeq2SeqLM
-
-print("➡️ Pre-downloading ONNX models...")
-WhisperProcessor.from_pretrained("distil-whisper/distil-large-v3")
-ORTModelForSeq2SeqLM.from_pretrained("distil-whisper/distil-large-v3.5-ONNX")
-print("✅ Pre-download complete.")
+from transformers import pipeline
+pipeline('automatic-speech-recognition', model='distil-whisper/distil-large-v3')
 PY
 
-# App
 COPY ./app.py /code/app.py
 
 EXPOSE 7860
-CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"]
+CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"]