Kyle Duffy

My apologies for not running the PR check before submitting! I've run them locally and patched the problem in commit d8ed691 — hopefully this will go through correctly.

@francoishernandez I use `-reset_optim 'all'` in my experiments so I did not encounter this issue—good catch. I've recreated the problem, which seems to occur when you train a frozen model...

also, I have not faced the apex/fp16 issue—not sure why that would happen

I also require freezing the encoder/decoder for one of my experiments. I'm currently adding the feature and planning to bundle into a PR!

## Update I've overridden `encoder_decoder.data_utils.group_parallel_data` to refrain from aggregating matching NL utterances in order to match the dataset sizes. This allows the eval logic to run, though as it stands,...

I am also looking for the same functionality. Perhaps support for the `!includes` directive is a worthy feature? Something as in this [stackoverflow example](https://stackoverflow.com/a/9577670/10620318) could be very useful.

> I have the same issue even after adjusting package.json node version to 16. I also have this issue. @psygo Can you clarify what you mean when you said you...

The same issue exists with GCP functions. Would love to see #9 go through.