Spaces:
Running
Running
remove dockerfile, revert to reqs.txt
Browse files- Dockerfile +0 -98
- requirements.txt +2 -2
Dockerfile
DELETED
@@ -1,98 +0,0 @@
|
|
1 |
-
# Use a CUDA base image without preinstalled cuDNN to avoid conflicts
|
2 |
-
FROM nvidia/cuda:12.3.2-devel-ubuntu22.04
|
3 |
-
|
4 |
-
# Set environment variables
|
5 |
-
ENV DEBIAN_FRONTEND=noninteractive
|
6 |
-
ENV PYTHONUNBUFFERED=1
|
7 |
-
ENV PYTHONDONTWRITEBYTECODE=1
|
8 |
-
ENV TF_FORCE_GPU_ALLOW_GROWTH=true
|
9 |
-
|
10 |
-
# Install system dependencies
|
11 |
-
RUN apt-get update && apt-get install -y --no-install-recommends \
|
12 |
-
git wget curl ca-certificates \
|
13 |
-
python3 python3-pip python3-dev \
|
14 |
-
ffmpeg libsm6 libxext6 libgl1-mesa-glx && \
|
15 |
-
apt-get clean && rm -rf /var/lib/apt/lists/*
|
16 |
-
|
17 |
-
# --- Download and install cuDNN 9.3.0 ---
|
18 |
-
# Download the archive directly from NVIDIA
|
19 |
-
RUN wget -O /tmp/cudnn-linux-x86_64-9.3.0.75_cuda12-archive.tar.xz https://developer.download.nvidia.com/compute/cudnn/redist/cudnn/linux-x86_64/cudnn-linux-x86_64-9.3.0.75_cuda12-archive.tar.xz && \
|
20 |
-
tar -xJvf /tmp/cudnn-linux-x86_64-9.3.0.75_cuda12-archive.tar.xz -C /tmp && \
|
21 |
-
cp -P /tmp/cudnn-linux-x86_64-9.3.0.75_cuda12-archive/cuda/include/cudnn*.h /usr/local/cuda/include && \
|
22 |
-
cp -P /tmp/cudnn-linux-x86_64-9.3.0.75_cuda12-archive/cuda/lib64/libcudnn* /usr/local/cuda/lib64 && \
|
23 |
-
chmod a+r /usr/local/cuda/include/cudnn*.h /usr/local/cuda/lib64/libcudnn* && \
|
24 |
-
rm -rf /tmp/cudnn-linux-x86_64-9.3.0.75_cuda12-archive.tar.xz /tmp/cudnn-linux-x86_64-9.3.0.75_cuda12-archive
|
25 |
-
|
26 |
-
# Set environment variables for CUDA/cuDNN libraries
|
27 |
-
ENV LD_LIBRARY_PATH=/usr/local/cuda/lib64:/usr/lib/x86_64-linux-gnu:${LD_LIBRARY_PATH}
|
28 |
-
ENV PATH=/usr/local/cuda/bin:${PATH}
|
29 |
-
ENV CUDA_VISIBLE_DEVICES=0
|
30 |
-
|
31 |
-
# Set working directory
|
32 |
-
WORKDIR /app
|
33 |
-
|
34 |
-
# Copy requirements and install Python dependencies
|
35 |
-
COPY requirements.txt /app/
|
36 |
-
RUN pip3 install --no-cache-dir --upgrade pip setuptools wheel
|
37 |
-
# Install TensorFlow GPU support (using version 2.15.0 here for compatibility)
|
38 |
-
RUN pip3 install --no-cache-dir tensorflow==2.15.0
|
39 |
-
# Install the remaining packages from requirements.txt (skip dependency resolution)
|
40 |
-
RUN pip3 install --no-cache-dir --no-deps -r requirements.txt
|
41 |
-
RUN pip3 install --no-cache-dir tensorflow-hub==0.14.0
|
42 |
-
RUN pip3 install --no-cache-dir opencv-python-headless opencv-contrib-python-headless
|
43 |
-
|
44 |
-
# Copy the application code
|
45 |
-
COPY . /app/
|
46 |
-
|
47 |
-
# Create a CPU fallback setup for TensorFlow
|
48 |
-
RUN echo 'import tensorflow as tf\n\
|
49 |
-
import os\n\
|
50 |
-
os.environ["TF_CPP_MIN_LOG_LEVEL"] = "2"\n\
|
51 |
-
\ndef setup_tensorflow():\n\
|
52 |
-
try:\n\
|
53 |
-
physical_devices = tf.config.list_physical_devices("GPU")\n\
|
54 |
-
if physical_devices:\n\
|
55 |
-
print(f"Found {len(physical_devices)} GPU(s)")\n\
|
56 |
-
for device in physical_devices:\n\
|
57 |
-
tf.config.experimental.set_memory_growth(device, True)\n\
|
58 |
-
print(f"Enabled memory growth for {device}")\n\
|
59 |
-
else:\n\
|
60 |
-
print("No GPU found. Running on CPU.")\n\
|
61 |
-
except Exception as e:\n\
|
62 |
-
print(f"Error setting up TensorFlow: {e}")\n\
|
63 |
-
print("Disabling GPU and falling back to CPU")\n\
|
64 |
-
os.environ["CUDA_VISIBLE_DEVICES"] = "-1"\n\
|
65 |
-
\n\
|
66 |
-
setup_tensorflow()\n' > /app/tf_setup.py
|
67 |
-
|
68 |
-
# Patch FILM.py to ensure proper GPU/CPU fallback, if the file exists
|
69 |
-
RUN if [ -f "/app/FILM.py" ]; then \
|
70 |
-
sed -i '1s/^/import tensorflow as tf\nfrom tf_setup import setup_tensorflow\n/' /app/FILM.py && \
|
71 |
-
sed -i '/def __init__/a\ # Check if GPU is disabled and use CPU if needed\n if "CUDA_VISIBLE_DEVICES" in os.environ and os.environ["CUDA_VISIBLE_DEVICES"] == "-1":\n print("GPU is disabled, using CPU for FILM")\n self._device = "/cpu:0"\n else:\n self._device = "/gpu:0"\n print(f"FILM will use device: {self._device}")' /app/FILM.py && \
|
72 |
-
sed -i '/def __call__/a\ with tf.device(self._device):' /app/FILM.py && \
|
73 |
-
sed -i 's/ result = self._model/ try:\n result = self._model/g' /app/FILM.py && \
|
74 |
-
sed -i '/result = self._model/a\ except Exception as e:\n print(f"Error during model inference: {e}, trying CPU fallback")\n with tf.device("/cpu:0"):\n result = self._model(inputs, training=False)' /app/FILM.py && \
|
75 |
-
sed -i '1s/^/import os\n/' /app/FILM.py; \
|
76 |
-
fi
|
77 |
-
|
78 |
-
# Create a startup script that checks CUDA/cuDNN status and launches Streamlit
|
79 |
-
RUN echo '#!/bin/bash\n\
|
80 |
-
set -e\n\
|
81 |
-
\n\
|
82 |
-
echo "CUDA libraries:"\n\
|
83 |
-
ldconfig -p | grep cuda\n\
|
84 |
-
echo "cuDNN libraries:"\n\
|
85 |
-
ldconfig -p | grep cudnn\n\
|
86 |
-
\n\
|
87 |
-
python3 -c "import tensorflow as tf; print(\\"Num GPUs Available: \\", len(tf.config.list_physical_devices(\\"GPU\\")))" || {\n\
|
88 |
-
echo "TensorFlow GPU test failed, falling back to CPU"\n\
|
89 |
-
export CUDA_VISIBLE_DEVICES=-1\n\
|
90 |
-
}\n\
|
91 |
-
\n\
|
92 |
-
exec streamlit run app.py --server.port=8501 --server.address=0.0.0.0\n' > /app/start.sh && chmod +x /app/start.sh
|
93 |
-
|
94 |
-
# Expose the port for Streamlit
|
95 |
-
EXPOSE 8501
|
96 |
-
|
97 |
-
# Use the startup script as the container's entrypoint
|
98 |
-
CMD ["/app/start.sh"]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
requirements.txt
CHANGED
@@ -14,6 +14,6 @@ torch
|
|
14 |
torchvision
|
15 |
lpips
|
16 |
# peft
|
17 |
-
|
18 |
-
|
19 |
opencv_python
|
|
|
14 |
torchvision
|
15 |
lpips
|
16 |
# peft
|
17 |
+
tensorflow==2.18.0
|
18 |
+
tensorflow_hub==0.16.1
|
19 |
opencv_python
|