google-research/text-to-text-transfer-transformer

Health v2 · maintenance only

Adoption not part of the score

6544 stars · 797 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-29, confidence not recorded

The official T5 library from Google Research, implementing the text-to-text transfer transformer for NLP tasks like summarization, question answering, and translation. It provides code for dataset loading, preprocessing, task mixing, fine-tuning, and evaluation of pre-trained T5 models.

Use cases

• fine-tune T5 on my own NLP task
• reproduce the T5 paper experiments
• preprocess and mix text-to-text training datasets
• download and run pre-trained T5 checkpoints
• train large seq2seq models on TPU
• evaluate models on summarization and question answering benchmarks

When to choose

• you want to fine-tune or evaluate the original T5 models in TensorFlow
• you need the exact dataset preprocessing and task-mixing code from the T5 paper
• you are doing research that builds on the T5 codebase

When to avoid

• you are starting fresh with T5 — use T5X (JAX/Flax) instead, as recommended by the authors
• you need actively developed features or modern LLM tooling
• you want a PyTorch-native implementation

Facets

library · maturity maintenance

machine-learning llm-training nlp etl machine-learning deep-learning large-language-models python cloud t5 transfer-learning text-to-text tensorflow sentencepiece pretrained-models research-code natural-language-processing gpu

Member repositories

For agents

markdown · JSON · MCP: product_card(name="google-research/text-to-text-transfer-transformer")