-
Notifications
You must be signed in to change notification settings - Fork 25.9k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[s2s] clean up + doc #6184
[s2s] clean up + doc #6184
Conversation
this is a follow up to huggingface#6149 - there was no need to add newly added options to finetune.sh - reverted that change - added a hint to users how to get all the options (--help)
Codecov Report
@@ Coverage Diff @@
## master #6184 +/- ##
==========================================
+ Coverage 78.29% 78.50% +0.20%
==========================================
Files 146 146
Lines 26607 26607
==========================================
+ Hits 20832 20887 +55
+ Misses 5775 5720 -55
Continue to review full report at Codecov.
|
examples/seq2seq/README.md
Outdated
To see all the possible command line options, run: | ||
|
||
```bash | ||
./finetune.sh --help |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
why not python finetune.py --help
? makes more sense to me.
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Either way works, but README.md uses the bash script in its usage examples:
https://github.com/huggingface/transformers/blob/master/examples/seq2seq/README.md#finetuning-training-params
Whatever you choose ideally let's make it consistent?
--encoder_layerdrop 0.1 \ | ||
--decoder_layerdrop 0.1 \ | ||
--dropout 0.1 \ | ||
--attention_dropout 0.1 \ |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Personally I would make dropout = 0.1
, and other types of dropout = 0 (but still list them here). Just a suggestion. @sshleifer
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
That resulted will be produced by BartConfig
:
encoder_layerdrop=0.0,
decoder_layerdrop=0.0,
attention_dropout=0.0,
dropout=0.1,
Co-authored-by: Sam Shleifer <sshleifer@gmail.com>
this is a follow up to #6149
@sshleifer