Update Dockerfile

Dockerfile

@@ -1,62 +1,61 @@
-# PyTorch + CUDA 12.1 + cuDNN 8 (passt zur L4)
+# Use PyTorch with CUDA 12.1 for L4 GPU compatibility
 FROM pytorch/pytorch:2.2.2-cuda12.1-cudnn8-runtime
 
-# -- System-Pakete minimal, aber was wir wirklich brauchen --
+# Install system dependencies
 RUN apt-get update && \
     apt-get install -y --no-install-recommends \
-        git-lfs build-essential portaudio19-dev ffmpeg && \
+        git \
+        git-lfs \
+        build-essential \
+        portaudio19-dev \
+        ffmpeg \
+        curl \
+        wget && \
     rm -rf /var/lib/apt/lists/*
 
-# Non-root-User (Spaces-Empfehlung)
+# Create non-root user for Hugging Face Spaces
 RUN useradd -m -u 1000 user
 USER user
 WORKDIR /app
+
+# Set environment variables
 ENV PATH="/home/user/.local/bin:$PATH"
+ENV HF_HOME=/app/.cache
+ENV PYTHONPATH="/app:$PYTHONPATH"
 
-# Orpheus-/SNAC-Code + Server
+# Copy application files
 COPY --chown=user . /app
 
-ENV HF_HOME=/app/.cache
-ENV VLLM_USE_LM_FORMAT_ENFORCER=0
-# GPU-freundliches Torch-Upgrade (falls gewünscht)
-#RUN pip install --no-cache-dir \
-#        torch==2.3.1+cu121 torchaudio==2.3.1+cu121 \
-#        --index-url https://download.pytorch.org/whl/cu121
+# Upgrade pip and install Python dependencies
+RUN pip install --upgrade pip
 
-RUN pip install --no-cache-dir "transformers==4.40.2" "lm-format-enforcer==0.9.8"
-RUN pip install --no-cache-dir vllm>=0.9.0
+# Install core dependencies first
+RUN pip install --no-cache-dir \
+    torch==2.2.2 \
+    transformers==4.40.2 \
+    vllm>=0.9.0
 
-# Python-Abhängigkeiten
+# Install requirements
 COPY --chown=user requirements.txt .
-RUN pip install --upgrade pip && \
-    pip install --no-cache-dir -r requirements.txt
+RUN pip install --no-cache-dir -r requirements.txt
 
-RUN pip install --no-cache-dir "realtimetts[system]>=0.5.5"
+# Install Orpheus TTS module
+RUN pip install --no-cache-dir -e ./orpheus-tts
 
-RUN pip install --no-cache-dir flashinfer-cu121-preview || echo "FlashInfer not available – continuing without."
-
-# nur *diese* Engine-Pflicht nachliefern, aber ohne Resolver:
-# RUN pip install --no-cache-dir pyttsx3==2.90 --no-deps
-
-# optional, um Warn-Spam zu reduzieren
-# RUN pip install --no-cache-dir azure-cognitiveservices-speech==1.33.0 --no-deps \
-#                 tqdm==4.66.1 --no-deps
+# Install additional optimizations for L4 GPU
+RUN pip install --no-cache-dir flash-attn --no-build-isolation || echo "Flash attention not available, continuing..."
 
+# Expose port for Hugging Face Spaces
 EXPOSE 7860
 
-# ───── Environment ───────────────────────────────────────
-
-ENV ORPHEUS_MODEL=SebastianBodza/Kartoffel_Orpheus-3B_german_synthetic-v0.1
-ENV MODEL_ID="SebastianBodza/Kartoffel_Orpheus-3B_german_synthetic-v0.1"
-ENV ORPHEUS_API_URL=http://127.0.0.1:1234
+# Environment variables for the application
+ENV PYTHONUNBUFFERED=1
+ENV UVICORN_HOST=0.0.0.0
+ENV UVICORN_PORT=7860
 
-# ───── Entrypoint ────────────────────────────────────────
-CMD bash -c "\
-  python -m vllm.entrypoints.openai.api_server \
-         --model ${MODEL_ID} \
-         --port 1234 \
-         --dtype bfloat16 \
-         --gpu-memory-utilization 0.85 \
-         --max-model-len 8192 & \ 
-  uvicorn app:app --host 0.0.0.0 --port 7860"
+# Health check using FastAPI health endpoint
+HEALTHCHECK --interval=30s --timeout=30s --start-period=60s --retries=3 \
+    CMD curl -f http://localhost:7860/health || exit 1
 
+# Start the FastAPI application with uvicorn
+CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860", "--workers", "1"]