Skip to content
This repository has been archived by the owner on Nov 3, 2023. It is now read-only.

[docs] Move a bunch of short options to long options. #3385

Merged
merged 3 commits into from
Jan 27, 2021

Conversation

stephenroller
Copy link
Contributor

Patch description
Global string replaces mostly, so some things may still be missing.

Still TODO: docstrings in parlai/scripts

Testing steps
CI render of website

Copy link
Contributor

@klshuster klshuster left a comment

Choose a reason for hiding this comment

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

sweet

@@ -364,34 +364,34 @@ Here are some examples:

```python
# Train a seq2seq model on the "10k training examples" bAbI task 1 with batch size of 32 examples until accuracy reaches 95% on validation (requires pytorch):
parlai train_model -t babi:task10k:1 -m seq2seq -mf /tmp/model_s2s -bs 32 -vtim 30 -vcut 0.95
parlai train_model --task babi:task10k:1 --model seq2seq --model-file /tmp/model_s2s --batchsize 32 --validation-every-n-secs 30 -vcut 0.95
Copy link
Contributor

Choose a reason for hiding this comment

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

--validation-cutoff (didn't even know this was an option)

docs/source/tutorial_basic.md Outdated Show resolved Hide resolved
```

Now, we'll train the model. This will take a while to reach convergence.

```bash
# train transformer ranker
parlai train_model -t twitter -mf /tmp/tr_twitter -m transformer/ranker -bs 16 -vtim 3600 -cands batch -ecands batch --data-parallel True
parlai train_model --task twitter --model-file /tmp/tr_twitter --model transformer/ranker --batchsize 16 --validation-every-n-secs 3600 --candidates batch --eval-candidates batch --data-parallel True
Copy link
Contributor

@mojtaba-komeili mojtaba-komeili Jan 16, 2021

Choose a reason for hiding this comment

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

validation-every-n-secs is too long. I know this is the doc, but is that possible to change that into something like validation-freq

Copy link
Contributor

Choose a reason for hiding this comment

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

that's a valid point (no pun intended) but the issue is that we have several (and soon to be more) ways of defining "frequency" (epochs, seconds, soon-to-be training steps).

perhaps we can change them all to --validation-freq-secs, --validation-freq-steps, --validation-freq-epochs

Copy link
Contributor Author

Choose a reason for hiding this comment

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

I'm hesitant to change the names of options that have been around 3 years :/

@stephenroller stephenroller merged commit 89603d6 into master Jan 27, 2021
@stephenroller stephenroller deleted the longopttutorial branch January 27, 2021 00:33
stephenroller added a commit that referenced this pull request Feb 11, 2021
* Move a bunch of short options to long options.

* Move some options in docstrings.

* A few missed things.
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants