-
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] add support for overriding config params #6149
Conversation
…p 0.1 --dropout 0.1 --attention_dropout 0.1
Codecov Report
@@ Coverage Diff @@
## master #6149 +/- ##
==========================================
+ Coverage 78.35% 79.68% +1.32%
==========================================
Files 146 146
Lines 26403 26403
==========================================
+ Hits 20689 21039 +350
+ Misses 5714 5364 -350
Continue to review full report at Codecov.
|
- optional - goes into model.config
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.
Can you add test cases:
- good case where model.config has the new value
- case (T5) that hits your assert. its called
dropout_rate
in t5. We can expose that or leave it for future work.
I'm surprised you didn't need to change the CHEAP_ARGS
constant in the tests.
The code looks perfect. |
because the new args are optional? Unless you mean something else. ...Working on the tests. |
Added tests as suggested. |
good alias sty () {
make style
flake8 examples templates tests src utils
} |
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)
add support for overriding model params:
as requested at #6018
README.md
seems to be mostly the editor removing superfluous whitespace - not sure why github shows it - normally it doesn't. The only added doc section is https://github.com/stas00/transformers/blob/seq2seq-train_params-1/examples/seq2seq/README.md#finetuning-training-params