microsoft/NeuronBlocks

Health v2 · maintenance only

Adoption not part of the score

1452 stars · 192 forks observed · 2026-08-28

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

NeuronBlocks is an NLP deep learning modeling toolkit from Microsoft that lets users build end-to-end neural network training and inference pipelines for NLP tasks using composable blocks and JSON configuration files. It includes a Block Zoo of reusable neural components and a Model Zoo of prebuilt models for tasks like classification, sequence labeling, and question answering matching.

Use cases

• build nlp deep learning models without writing code
• train a text classifier with config
• do sequence labeling like slot tagging
• train a question answering matching model
• compress models via knowledge distillation
• prototype neural architectures from reusable blocks

When to choose

• you want config-driven NLP model training in PyTorch
• you need prebuilt models for common NLP tasks like classification or slot tagging
• you want to experiment with model architectures by composing blocks

When to avoid

• you need state-of-the-art transformer fine-tuning with modern ecosystems like Hugging Face
• you require active community support or frequent updates
• you work outside NLP or need non-PyTorch frameworks

Facets

framework · maturity maintenance

machine-learning deep-learning nlp llm-training deep-learning machine-learning python windows cross-platform pytorch model-zoo block-zoo -configuration knowledge-distillation model-compression sequence-labeling text-classification question-answering natural-language-processing linux

Member repositories

For agents

markdown · JSON · MCP: product_card(name="microsoft/NeuronBlocks")