Example how to pretrain lm + introduction of config_name #57
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
So one can pretrain a language model from commandline The limit was added to support quick tests
| @@ -280,7 +288,7 @@ def train_(self, dataset_or_path, tokenizer=None, **train_config): | |||
| print("Language model saved to", self.experiment_path) | |||
|
|
|||
| def validate(self): | |||
| raise NotImplementedError("The validation on the language model is not implemented.") | |||
| return "not implemented" | |||
There was a problem hiding this comment.
Do we really just want to return a string here?
| From command line: | ||
| ``` | ||
| $ bash prepare_wiki.sh de | ||
| $ python -W ignore -m multifit new multifit_paper_version replace_ --name my_lm - train_ --pretrain-dataset data/wiki/de-100 |
There was a problem hiding this comment.
Looks like there's a superfluous space between - and train-. Why do we use train_ here? What is the difference between train_ and train?
sebastianruder
approved these changes
Nov 17, 2019
There was a problem hiding this comment.
Hi Piotr, thanks for adding this. Looks good in general. I've added a few comments about minor things. In general, do you think it'd be possible to add a few short docstrings to explain things like bs, bptt, limit in load_lm_databunch for people not familiar with the library?
Closed
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
I've added ability to limit training set so we can use a test configuration 'multifit_mini_test` that executes in ~20 secs to test that the scripts are working.
Why config_name?
I've added it so we can know what training parameters we should load for the finetune-lm and and classifier. This parameters aren't stored along with a language model, only parameters used to build that model are saved.