Requirements for Vid2Seq

[lawrenceztang]

I'm having a lot of trouble running Vid2Seq. Here are my steps:

1. Create new TPU v2-8
2. gcloud compute tpus tpu-vm ssh node-1 --zone=us-central1-f
3. Python version is 3.8, so update Python 3.10. Install dependencies:

sudo apt update; sudo apt install build-essential libssl-dev zlib1g-dev \
libbz2-dev libreadline-dev libsqlite3-dev curl \
libncursesw5-dev xz-utils tk-dev libxml2-dev libxmlsec1-dev libffi-dev liblzma-dev

4. Install Python 3.10:

curl https://pyenv.run | bash
echo 'export PATH="$HOME/.pyenv/bin:$PATH"' >> ~/.bashrc
echo 'eval "$(pyenv init -)"' >> ~/.bashrc
echo 'eval "$(pyenv virtualenv-init -)"' >> ~/.bashrc
source ~/.bashrc
pyenv install 3.10.7
pyenv global 3.10.7

5. Pip can't find t5x so install t5x:

git clone --branch=main https://github.com/google-research/t5x
cd t5x
python3 -m pip install -e '.[tpu]' -f \
https://storage.googleapis.com/jax-releases/libtpu_releases.html

6. git clone https://github.com/google-research/scenic
7. pip install -r scenic/projects/vid2seq/requirements.txt
8. I get ERROR: Cannot install flax==0.5.0 and t5x==0.0.0 because these package versions have conflicting dependencies.

Now what do I do?

[peiliu0408]

Could you provide an update? I've been encountering numerous difficulties while using the video2seq codebase for inference.

[lawrenceztang]

Could you provide an update? I've been encountering numerous difficulties while using the video2seq codebase for inference.

I don't have any updates yet. I might try again sometime soon. Please let me know if you figure anything out too.