awslabs/sockeye

Health v2 · maintenance only

Adoption not part of the score

1215 stars · 320 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

Sockeye is an open-source sequence-to-sequence framework for Neural Machine Translation built on PyTorch, with distributed training and optimized inference. It powers Amazon Translate and is now in maintenance mode with no new features.

Use cases

• train a neural machine translation model on parallel corpora
• train a transformer seq2seq model with PyTorch
• run optimized inference for translation models
• fine-tune an encoder-decoder model for sequence transduction
• convert old MXNet Sockeye models to PyTorch
• score translations with a trained NMT model

When to choose

• you need a proven, production-grade NMT training and inference toolkit
• you want transformer-based seq2seq training with distributed support
• you need to run inference with existing Sockeye models

When to avoid

• you need actively developed features or new model architectures
• you want a general-purpose LLM or chat model framework
• you need long-term support for new PyTorch versions

Facets

library · maturity maintenance

machine-learning deep-learning llm-training machine-learning deep-learning python cross-platform neural-machine-translation seq2seq transformer pytorch encoder-decoder natural-language-processing gpu linux

Member repositories

For agents

markdown · JSON · MCP: product_card(name="awslabs/sockeye")