search

Shivanandroy / simpleT5

simpleT5 is built on top of PyTorch-lightning⚡️ and Transformers🤗 that lets you quickly train your T5 models.
MIT License
382 stars 61 forks source link

Adding logger #22

Closed versae closed 2 years ago

versae commented 2 years ago

With this change, users can provide a PyTorch Lightning logger object to the .train() method:

from pytorch_lightning.loggers import WandbLogger

wandb_logger = WandbLogger(project="my-project", name="run-name")

model.train(
    train_df=train_df,
    eval_df=eval_df,
    logger=wandb_logger
)
eliasws commented 2 years ago

@versae Seems like this got integrated without the PR. @Shivanandroy Can you make a new release?

Shivanandroy commented 2 years ago

@versae My bad! Apologies. I didn't see the PR, I was not actively managing it for couple of months. @eliasws It's already under pre-release. you can update simplet5 pip install -U simplet5

versae commented 2 years ago

All good 🤠