git clone https://github.com/wxj77/MeetingSummarization.git
pip install -r requirements.txt
python setup.py install
git add .
git commit -m 'I did what'
git push origin branch_name
...
- build docker image
sudo docker build . -t hmnet
- clone HMNet into MeetingSummarization:
git clone https://github.com/microsoft/HMNet
- download
model.pt
intoHMNet/ExampleInitModel/HMNet-pretrained
- run docker with gpu and attach volume:
sudo docker run --gpus all -v /home/wji/language/MeetingSummarization:/MeetingSummarization -it hmnet /bin/bash
- test nvidia:
nvidia-smi
- test cuda:
python -c 'import torch; print(torch.cuda.is_available());'
- go into HMNet directory: cd /MeetingSummarization/HMNet
- run train script with 1 gpu:
CUDA_VISIBLE_DEVICES="0" mpirun -np 1 --allow-run-as-root python PyLearn.py train ExampleConf/conf_hmnet_AMI