Skip to content

This repo contains a list of stand alone scripts for running torchbench models one by one.

License

Notifications You must be signed in to change notification settings

pytorch-tpu/run_torchbench

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

46 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

run_torchbench

This repo contains a list of stand alone scripts for running torchbench models one by one.

Dependencies

1. Torch XLA2:

Please install following the instructions in https://github.com/pytorch/xla/tree/master/experimental/torch_xla2

2. TorchBench

Important: Please git clone torchbench in this directory and install from source

pip install torchvision torchaudio
git clone https://github.com/pytorch/benchmark.git
cd benchmark
python3 install.py

NOTE: python install.py will install ALL models. If you are only interested in running one model, you can install it via python install.py model_name this way it will be much faster.

NOTE:* if you met error like ERROR: Cannot install -r requirements.txt (line 12) because these package versions have conflicting dependencies..., please try to reinstall env with:

pip3 uninstall torch torchvision torchaudio -y
pip3 install --pre torch torchvision torchaudio --index-url https://download.pytorch.org/whl/nightly/cu121

3. Run one torch bench model under torch xla2

python models/name_of_the_model.py

For example,

python models/BERT_pytorch.py

The scripts are identical per model except the model name. The goal is for people to add loggings and edit the script as they see fit. And able to focus just one model.

About

This repo contains a list of stand alone scripts for running torchbench models one by one.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 3

  •  
  •  
  •  

Languages