# pytorch/glow

Compiler for Neural Network hardware accelerators

Repository: https://github.com/pytorch/glow
Canonical: https://ross.abutalabs.com/products/pytorch-glow
Language: C++
License: Apache-2.0
License Family: permissive
Archived: true
Last push: 2024-05-11T13:35:10+00:00

## Health v2 (maintenance only)
Score: 10/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 35, longevity 100
- inputs: {"age_days": 3260, "days_push": 844, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, archived
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3320, forks 701 (observed 2026-08-28T04:07:56.269784+00:00)

## What it is
Glow is a machine learning compiler and execution engine for neural network hardware accelerators, developed under the PyTorch project. It lowers neural network dataflow graphs through a two-phase strongly-typed IR to perform domain-specific optimizations, memory optimizations, and machine-specific code generation for hardware backends.

## Use cases
- compile neural network graphs for custom hardware accelerators
- build a backend for a machine learning framework targeting specialized chips
- optimize memory allocation and instruction scheduling for NN inference
- port deep learning models to embedded AI hardware
- experiment with ML compiler IR design and lowering techniques

## When to choose
- you are a hardware vendor building a compiler backend for an AI accelerator
- you need to lower ONNX-style neural network graphs to custom hardware
- you want state-of-the-art graph- and memory-level optimizations for NN execution

## When to avoid
- you just need to run inference on CPUs or GPUs with an existing runtime like PyTorch or ONNX Runtime
- you need a actively developed tool - development has slowed and the project is largely in maintenance
- you want a high-level training framework rather than a compiler backend

## Facets
- artifact type: library
- maturity: maintenance
- function: compiler, machine-learning, llm-inference
- domain: machine-learning, compilers, gpu-computing, deep-learning
- platform: cpp
- tags: neural-network-compiler, hardware-accelerators, code-generation, execution-engine, ml-compiler, linux, macos

## Member repositories
- pytorch/glow (main) score 10

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:07:56.269784+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 2026-08-29T18:41:51.295524+00:00, confidence not recorded.
  - readme: https://github.com/pytorch/glow (fetched 2026-08-28T04:07:56.269784+00:00, sha 8d96b4f0913b)
- Data as of 2026-08-30T08:39:29.467469+00:00.