# ML-GSAI/LLaDA

Official PyTorch implementation for "Large Language Diffusion Models"

Repository: https://github.com/ML-GSAI/LLaDA
Canonical: https://ross.abutalabs.com/products/llada
Language: Python
License Family: other
Last push: 2026-07-15T08:24:01+00:00

## Health v2 (maintenance only)
Score: 62/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 92, release rhythm 35, longevity 40
- inputs: {"age_days": 566, "days_push": 49, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3943, forks 276 (observed 2026-08-28T04:08:30.228654+00:00)

## What it is
Official PyTorch implementation of LLaDA, a family of large language diffusion models (8B base/instruct, MoE, and iLLaDA variants) with pretrained weights on Hugging Face. Includes inference code, batch inference, and lm-evaluation-harness-based evaluation.

## Use cases
- run inference with a diffusion-based language model
- evaluate a diffusion LLM on standard benchmarks
- experiment with masked diffusion text generation
- try an alternative to autoregressive LLMs
- fine-tune or study large language diffusion models
- generate text with an 8B diffusion model in PyTorch

## When to choose
- you want to research or experiment with diffusion-based language models
- you need official weights and evaluation code for LLaDA
- you want a non-autoregressive alternative to transformer LLMs

## When to avoid
- you need a production-ready LLM serving stack
- you lack GPU resources for 8B-scale models
- you need a permissively licensed model for commercial use (no license specified)

## Facets
- artifact type: library
- maturity: active
- function: llm-inference, machine-learning, deep-learning, llm-training, benchmarking
- domain: large-language-models, deep-learning, artificial-intelligence
- platform: python
- tags: diffusion-language-model, pytorch, research-code, masked-diffusion, huggingface, natural-language-processing, gpu, linux

## Member repositories
- ML-GSAI/LLaDA (main) score 62

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:30.228654+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:24:36.258453+00:00, confidence not recorded.
  - readme: https://github.com/ML-GSAI/LLaDA (fetched 2026-08-28T04:08:30.228654+00:00, sha 5a5952d82e4e)
  - registry_pypi: https://pypi.org/pypi/llada/json (fetched 2026-08-29T09:18:12.953241+00:00, sha 1ba1a82befc8)
- Data as of 2026-08-30T08:39:29.467469+00:00.