Skip to content

[docs] [testing] distributed training#7993

Merged
sgugger merged 4 commits intohuggingface:masterfrom
stas00:test-distrib
Oct 26, 2020
Merged

[docs] [testing] distributed training#7993
sgugger merged 4 commits intohuggingface:masterfrom
stas00:test-distrib

Conversation

@stas00
Copy link
Copy Markdown
Contributor

@stas00 stas00 commented Oct 22, 2020

We figured out how to support distributed training with pytest, this is a preliminary doc snippet to help those in need to find the current implementation. I'm sure it will evolve as we have more tests with varying needs, but for now that's all we have.

@sgugger

Copy link
Copy Markdown
Collaborator

@sgugger sgugger left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM, thanks!

@sgugger
Copy link
Copy Markdown
Collaborator

sgugger commented Oct 26, 2020

One followup is to update the test_trainer_distributed to work with pytest. Then ideally, if we could have one command to run all those tests, that would be awesome (maybe we can use a pytest marker to mark all distributed-specific tests so it's easy to select them all?)

@sgugger sgugger merged commit 101186b into huggingface:master Oct 26, 2020
@stas00 stas00 deleted the test-distrib branch October 26, 2020 17:19
@stas00
Copy link
Copy Markdown
Contributor Author

stas00 commented Oct 26, 2020

I will port test_trainer_distributed - thank you for flagging that, @sgugger
Tracking it here: #8058

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants