Raphael Tang
Raphael Tang
The ConvRNN implementation doesn't use anything in `common.*`, not to mention that `getData.sh` doesn't work -- and it's bad practice, since we agreed to use `Castor-data` only. Related: #102
@achyudhk reports that CharCNN on some dataset uses 63GB of RAM (Hydra and Dragon both have 64GB). I think a solution would be some mechanism for moving data between disk...
Currently, we're using torchtext 0.2.* -- we should update to the next major version, 0.3.
Can you freely open source the code so that I can incorporate it into another noncommercial open source project? The MIT license works. I'll add an attribution section about citing...
We plan to wrap up the existing codebase into a historical package and then overhaul it.
Make the `T5CachedModelLoader` class more general to include *BERT -- should just be a few lines.
The latest `transformers` version (4.50.0) results in some cryptic error. These dependency specifiers should work: ``` transformers~=4.49.0 accelerate>=0.26.0 ```
Add missing [PlaybookType](https://cloud.google.com/dialogflow/cx/docs/reference/rest/v3beta1/projects.locations.agents.playbooks#playbooktype) to the playbook proto so that users can CREATE and PATCH routine playbooks. Currently, Google's API SDKs do not expose this field, limiting users to operations on...