Trainer tool finalized and completed!
Published by gagan3012 over 3 years ago
Val acc added
Published by gagan3012 over 3 years ago
Bug fixes
Published by gagan3012 over 3 years ago
New Upload to hf hub module
Published by gagan3012 over 3 years ago
Documentation updated along with sematic versioning
Published by gagan3012 over 3 years ago
This version features a tested trainer which can be used in 4 lines of code:
from keytotext import KeytotextTrainer
model = KeytotextTrainer()
model.from_pretrained(model_name="t5-small")
model.train(data_df=df,batch_size=4, max_epochs=3, use_gpu=True)
model.save_model()
Published by gagan3012 over 3 years ago
This release features the new Trainer module
More details coming soon
Published by gagan3012 over 3 years ago
Changes:
- Bug Fixes
- Maintaining new models
Published by gagan3012 over 3 years ago
Changes:
- Refactoring of code
- Ability to add new models too
Published by gagan3012 over 3 years ago
v0.2.3 :
- Bug fixes
- New models added
Published by gagan3012 over 3 years ago
Changes:
- Now keytotext supports new models trained by other people too
- A new fine-tuning script
Published by gagan3012 over 3 years ago
Bug fixes
Published by gagan3012 over 3 years ago
Latest Release:
- Completed API
- Completed testing
- completed all Evals
- UI Improvements too
Published by gagan3012 over 3 years ago
Changes:
- Updates to Eval pipeline
Published by gagan3012 over 3 years ago
Changes:
- Added Trainer API
- Added Eval pipeline
Published by gagan3012 over 3 years ago
Latest release
Published by gagan3012 over 3 years ago
Updates
Published by gagan3012 over 3 years ago
Production release- 0.1.0
