https://github.com/himkt/lstm_benchmark

https://github.com/himkt/lstm_benchmark

Last synced: 10 months ago
JSON representation

• Host: GitHub
• URL: https://github.com/himkt/lstm_benchmark
• Owner: himkt
• Created: 2018-10-08T08:41:25.000Z (over 7 years ago)
• Default Branch: master
• Last Pushed: 2018-10-08T10:47:55.000Z (over 7 years ago)
• Last Synced: 2025-01-13T05:08:45.067Z (over 1 year ago)
• Language: Python
• Size: 13.7 KB
• Stars: 0
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

README

          
## Benchmark for Chainer's LSTM and PyTorch's LSTM

### GPU

> python benchmark_lstms.py --gpuid 0 --trial 100

https://github.com/himkt/lstm_benchmark/blob/master/result/GPU.summary.md

### CPU

> python benchmark_lstms.py --gpuid -1 --trial 100

https://github.com/himkt/lstm_benchmark/blob/master/result/CPU.summary.md

It looks PyTorch's LSTM is significantly slow.

PyTorch's LSTM [ellapsed times](https://github.com/himkt/lstm_benchmark/blob/master/result/lstm_pytorch_CPU.time.txt) seems to be unstable.