Transformers

Use it when

• load, fine-tune, and evaluate pretrained transformer models
• use a common API across text vision audio and multimodal model families

What it solves

Not the fit when

• model weights and datasets have their own licenses and resource needs

Install

pip install transformers

Invoke

Load a version-pinned tokenizer and model with `from_pretrained`, then use a pipeline, Trainer, or explicit PyTorch loop and retain the model revision.

Alternatives

No reviewed alternatives recorded yet.