Happy Transformer

Documentation and news: happytransformer.com

Join our Discord server:

Happy Transformer makes it easy to fine-tune and perform inference with NLP Transformer models.

3.0.0

1. DeepSpeed for training
2. Apple's MPS for training and inference
3. WandB to track training runs
4. Data supplied for training is automatically split into portions for training and evaluating
5. Push models directly to Hugging Face's Model Hub

Read about the full 3.0.0 update including breaking changes here.

Tasks

Tasks	Inference	Training
Text Generation	✔	✔
Text Classification	✔	✔
Word Prediction	✔	✔
Question Answering	✔	✔
Text-to-Text	✔	✔
Next Sentence Prediction	✔
Token Classification	✔

Quick Start

pip install happytransformer

from happytransformer import HappyWordPrediction
#--------------------------------------#
happy_wp = HappyWordPrediction()  # default uses distilbert-base-uncased
result = happy_wp.predict_mask("I think therefore I [MASK]")
print(result)  # [WordPredictionResult(token='am', score=0.10172799974679947)]
print(result[0].token)  # am

Maintainers

• Eric Fillion Lead Maintainer
• Ted Brownlow Maintainer

Tutorials

Text generation with training (GPT-Neo)

Text classification (training)

Text classification (hate speech detection)

Text classification (sentiment analysis)

Word prediction with training (DistilBERT, RoBERTa)

Top T5 Models

Grammar Correction

Fine-tune a Grammar Correction Model