Rerender_A_Video
Rerender_A_Video copied to clipboard
Black frames between key frames after full video translation
I run project on RTX 3090 TI, in docker container (dockerfile below)
FROM nvcr.io/nvidia/cuda:11.3.1-cudnn8-runtime-ubuntu20.04
ENV DEBIAN_FRONTEND=noninteractive
# install base dependencies
RUN apt update && apt install curl wget nano libgl1 -y
# install python3.8+pip
RUN apt install software-properties-common -y && \
add-apt-repository ppa:deadsnakes/ppa -y && \
apt update && \
apt install python3.8 python3.8-distutils -y && \
curl -sS https://bootstrap.pypa.io/get-pip.py | python3.8
WORKDIR /app
RUN pip install numpy==1.23.1 torch==1.12.1+cu113 torchvision==0.13.1+cu113 torchaudio==0.12.1 --extra-index-url https://download.pytorch.org/whl/cu113
COPY requirements.txt .
RUN pip install -r requirements.txt
RUN apt update && apt install build-essential cuda-toolkit-11-3 ffmpeg -y
COPY deps/ /app/deps/
COPY install.py .
RUN python3.8 install.py
CMD [ "/bin/bash" ]
and fix build-linux-cpu+cuda.sh
replaced nvcc -arch compute_50 ...
with nvcc -arch compute_86 ...
After run propogation, i have video with black frames between keyfames (screenshot below)
What could be the problem?