Text Generation Pipeline

Resolves #220 by using a TextGenerationPipeline. It also improves consistency as all most of the other tasks use a Pipeline (other than next sentence prediction).

2.2.1: Updated Minimum Version for Transformers

Increased the minimum version for Hugging Face's transformers library from 4.0.0 to 4.4.0.

See #215 for more details

2.2.0: Text Generation With Training + More

New Features:

Text generation with training
Word prediction training
Saving/loading models
Saving/loading preprocessed data
You can now change the batch size when training and evaluating
Dataclasses can now be used for all finetuning “arg” parameters

2.1.0: New Models and Token Classification

• You can now use any model that's available on Hugging Face's model distribution network for text classification, word prediction, next sentence prediction, question answering and token classification.
• You can now perform token classification using HappyTokenClassification

Addresses:
#193
#133
#121

2.0.0: Fully Redesigned

Introducing Version 2.0.0!

We fully redesigned Happy Transformer from the ground up.

New Features:
Question answering training
Multi label text classification training
Single predictions for text classification

Deprecated Features:
Masked word prediction training
Masked word prediction with multiple masks

Breaking changes:
Everything

Happy Transformer have been redesigned to promote scalability. Now it's easier than ever to add new models and features, and we encourage you to create PRs to contribute to the project.

1.1.3

• Cleaned up logging
• Resolved #137
• Added "use_probability" parameter to next sentence prediction methods
• Resolved #144
• Fixed a bug that prevented using a GPU for masked word prediction

1.1.2

Fixed issues regarding fine tuning masked word prediction models.

1.1.1-alpha

Updated version to 1.1.1

1.1.0-alpha

Updated setup.py to 1.1.1-alpha

PyPi 1.1.0

Update setup.py