IBM/pytorch-seq2seq

Health v2 · maintenance only

Flags: archived

Adoption not part of the score

1515 stars · 367 forks observed · 2026-08-28

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

An open-source framework for building sequence-to-sequence (seq2seq) models in PyTorch, with modular and extensible components for models, training, inference, and checkpoints. It targets research and experimentation on encoder-decoder architectures such as machine translation and conversational models.

Use cases

• train a seq2seq model for machine translation
• build an encoder-decoder model with attention in PyTorch
• experiment with sequence-to-sequence architectures
• train a conversational chatbot model
• use pre-trained word embeddings in a seq2seq model
• benchmark seq2seq models on WMT translation tasks

When to choose

• you want a modular, extensible PyTorch framework specifically for seq2seq models
• you are researching or teaching encoder-decoder architectures
• you need a lightweight library rather than a full training platform

When to avoid

• you need modern transformer implementations with active development
• you want a high-level framework like Hugging Face Transformers with pretrained models
• you need production-grade NLP pipelines rather than research code

Facets

framework · maturity maintenance

machine-learning deep-learning machine-learning python seq2seq pytorch encoder-decoder attention machine-translation natural-language-processing

Member repositories

For agents

markdown · JSON · MCP: product_card(name="IBM/pytorch-seq2seq")