Fix parallel load non parallel model - Githubissues

amjltc295 / pytorch-golden-template

PyTorch Golden Template (under development)

13 stars 4 forks source link

Fix parallel load non parallel model #49

Closed amjltc295 closed 5 years ago

amjltc295 commented 5 years ago

Why do we need this PR?

Fix #48

How is it implemented?

Load model.module weight when multiple GPUs are used

Other changes

Change lr decay rate to 0.9

PR readiness checklist

[x] Did it pass the Flake8 check?

[x] Did you test this PR?