up

Dockerfile

@@ -1,27 +1,53 @@
-FROM python:3.11
+# Stage 1: Get uv installer
+FROM ghcr.io/astral-sh/uv:0.2.12 as uv
 
-# Set the working directory to /code
-WORKDIR /code
+# Stage 2: Main application image
+FROM python:3.11.9-slim-bookworm
 
-# Copy the current directory contents into the container at /code
-COPY ./requirements.txt /code/requirements.txt
+# Copy uv from first stage
+COPY --from=uv /uv /uv
 
-# Install requirements.txt 
-RUN pip install --no-cache-dir --upgrade -r /code/requirements.txt
+# Create virtual environment with uv
+RUN --mount=type=cache,target=/root/.cache/uv \
+    /uv venv /opt/venv
 
-# Set up a new user named "user" with user ID 1000
-RUN useradd -m -u 1000 user
+# Set environment variables
+ENV VIRTUAL_ENV=/opt/venv \
+    PATH="/opt/venv/bin:$PATH"
 
-# Switch to the "user" user
+# Create user and set permissions (required for HF Spaces)
+RUN useradd -m -u 1000 user && \
+    chown -R user /opt/venv
+
+# Switch to user context
 USER user
-# Set home to the user's home directory
+WORKDIR /app
+
+# Set home to user's home directory
 ENV HOME=/home/user \
-	PATH=/home/user/.local/bin:$PATH
+    PATH=/home/user/.local/bin:$PATH \
+    HF_HOME=/home/user/.cache/huggingface \
+    UV_CACHE_DIR=/app/.uv-cache
+
+# Create cache directory with proper permissions
+RUN mkdir -p $UV_CACHE_DIR && chown -R user:user $UV_CACHE_DIR
+
+# Copy requirements first for caching
+COPY --chown=user requirements.txt .
+
+# Install Python packages with uv caching
+RUN --mount=type=cache,target=$UV_CACHE_DIR,uid=1000,gid=1000 \
+    /uv pip install -r requirements.txt
+
+# Copy application code
+COPY --chown=user . .
 
-# Set the working directory to the user's home directory
-WORKDIR $HOME/app
+# Expose FastRTC port (matches HF Spaces default)
+EXPOSE 7860
 
-# Copy the current directory contents into the container at $HOME/app setting the owner to the user
-COPY --chown=user . $HOME/app
+# Check gpu and cuda
+RUN python -c "import torch; print(torch.cuda.is_available())"
+RUN nvidia-smi
 
+# Start the application using uvicorn (FastAPI)
 CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "7860"]

Dockerfile.pip

@@ -0,0 +1,27 @@
+FROM python:3.11.9
+
+# Set the working directory to /code
+WORKDIR /code
+
+# Copy the current directory contents into the container at /code
+COPY ./requirements.txt /code/requirements.txt
+
+# Install requirements.txt 
+RUN pip install --no-cache-dir --upgrade -r /code/requirements.txt
+
+# Set up a new user named "user" with user ID 1000
+RUN useradd -m -u 1000 user
+
+# Switch to the "user" user
+USER user
+# Set home to the user's home directory
+ENV HOME=/home/user \
+	PATH=/home/user/.local/bin:$PATH
+
+# Set the working directory to the user's home directory
+WORKDIR $HOME/app
+
+# Copy the current directory contents into the container at $HOME/app setting the owner to the user
+COPY --chown=user . $HOME/app
+
+CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "7860"]

Dockerfile.uv

@@ -1,49 +0,0 @@
-# Stage 1: Get uv installer
-FROM ghcr.io/astral-sh/uv:0.2.12 as uv
-
-# Stage 2: Main application image
-FROM python:3.10.12-slim-bookworm
-
-# Copy uv from first stage
-COPY --from=uv /uv /uv
-
-# Create virtual environment with uv
-RUN --mount=type=cache,target=/root/.cache/uv \
-    /uv venv /opt/venv
-
-# Set environment variables
-ENV VIRTUAL_ENV=/opt/venv \
-    PATH="/opt/venv/bin:$PATH"
-
-# Create user and set permissions (required for HF Spaces)
-RUN useradd -m -u 1000 user && \
-    chown -R user /opt/venv
-
-# Switch to user context
-USER user
-WORKDIR /app
-
-# Set home to user's home directory
-ENV HOME=/home/user \
-    PATH=/home/user/.local/bin:$PATH \
-    HF_HOME=/home/user/.cache/huggingface \
-    UV_CACHE_DIR=/app/.uv-cache
-
-# Create cache directory with proper permissions
-RUN mkdir -p $UV_CACHE_DIR && chown -R user:user $UV_CACHE_DIR
-
-# Copy requirements first for caching
-COPY --chown=user requirements.txt .
-
-# Install Python packages with uv caching
-RUN --mount=type=cache,target=$UV_CACHE_DIR,uid=1000,gid=1000 \
-    /uv pip install -r requirements.txt
-
-# Copy application code
-COPY --chown=user . .
-
-# Expose FastRTC port (matches HF Spaces default)
-EXPOSE 7860
-
-# Start the application using uvicorn (FastAPI)
-CMD ["uvicorn", "main:app", "--host", "0.0.0.0", "--port", "7860"]

main.py

@@ -99,20 +99,20 @@ stream = Stream(
             # If, after the user started speaking, there is a chunk with less than speech_threshold seconds of speech, the user stopped speaking. (default 0.1)
             speech_threshold=0.1,
         ),
-        model_options=SileroVadOptions(
-            # Threshold for what is considered speech (default 0.5)
-            threshold=0.5,
-            # Final speech chunks shorter min_speech_duration_ms are thrown out (default 250)
-            min_speech_duration_ms=250,
-            # Max duration of speech chunks, longer will be split (default float('inf'))
-            max_speech_duration_s=30,
-            # Wait for ms at the end of each speech chunk before separating it (default 2000)
-            min_silence_duration_ms=2000,
-            # Chunk size for VAD model. Can be 512, 1024, 1536 for 16k s.r. (default 1024)
-            window_size_samples=1024,
-            # Final speech chunks are padded by speech_pad_ms each side (default 400)
-            speech_pad_ms=400,
-        ),
+        #model_options=SileroVadOptions(
+        #    # Threshold for what is considered speech (default 0.5)
+        #    threshold=0.5,
+        # Final speech chunks shorter min_speech_duration_ms are thrown out (default 250)
+        #    min_speech_duration_ms=250,
+        # Max duration of speech chunks, longer will be split (default float('inf'))
+        #    max_speech_duration_s=30,
+        # Wait for ms at the end of each speech chunk before separating it (default 2000)
+        #    min_silence_duration_ms=2000,
+        # Chunk size for VAD model. Can be 512, 1024, 1536 for 16k s.r. (default 1024)
+        #    window_size_samples=1024,
+        # Final speech chunks are padded by speech_pad_ms each side (default 400)
+        #    speech_pad_ms=400,
+        #),
     ),
     # send-receive: bidirectional streaming (default)
     # send: client to server only
@@ -123,7 +123,8 @@ stream = Stream(
         gr.Textbox(label="Transcript"),
     ],
     additional_outputs_handler=lambda current, new: current + " " + new,
-    rtc_configuration=get_rtc_credentials(provider="hf") if os.getenv("APP_MODE") == "deployed" else None
+    rtc_configuration=get_rtc_credentials(provider="hf") if os.getenv("APP_MODE") == "deployed" else None,
+    concurrency_limit=6
 )
 
 app = FastAPI()

requirements.txt

@@ -1,9 +1,9 @@
-accelerate==1.4.0
-fastrtc==0.0.10
+accelerate
+fastrtc
 fastrtc[vad]
-python-dotenv==1.0.1
-transformers==4.49.0
-torch==2.6.0
-torchaudio==2.6.0
+python-dotenv
+transformers
+torch
+torchaudio
 fastapi
 uvicorn[standard]