# Jittor/jittor

Jittor is a high-performance deep learning framework based on JIT compiling and meta-operators.

Repository: https://github.com/Jittor/jittor
Canonical: https://ross.abutalabs.com/products/jittor
Homepage: https://cg.cs.tsinghua.edu.cn/jittor/
Language: Python
License: Apache-2.0
License Family: permissive
Topics: deep-learning, gpu, python, jittor, cuda
Last push: 2026-08-26T20:13:48+00:00

## Health v2 (maintenance only)
Score: 67/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 8, longevity 100
- inputs: {"age_days": 2359, "days_push": 7, "days_rel": 401, "gap_med": null, "n_releases_24m": 1}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3229, forks 324 (observed 2026-08-28T04:07:50.095716+00:00)

## What it is
Jittor is a high-performance deep learning framework from Tsinghua University based on just-in-time (JIT) compilation and meta-operators, with a Python front-end and CUDA/C++ back-end. It includes an integrated op compiler and tuner plus model libraries for vision, detection, segmentation, GANs, point clouds, and LLM inference.

## Use cases
- train neural networks with a pytorch-like api
- run deep learning models with jit-compiled optimized ops
- run large language model inference on limited hardware
- train image segmentation and detection models
- do differentiable rendering research
- run deep learning on gpu with automatic op tuning

## When to choose
- you want a pytorch-style dynamic graph framework with jit-compiled performance
- you need to run LLMs on modest hardware via JittorLLMs
- you want to customize operators without writing low-level code
- you work in computer vision research areas covered by its model libraries

## When to avoid
- your team and ecosystem depend on pytorch or tensorflow
- you need broad multi-vendor hardware support beyond cuda/cpu
- you require a large community and third-party library ecosystem
- you need windows-first or mobile deployment support

## Facets
- artifact type: framework
- maturity: active
- function: deep-learning, machine-learning, llm-inference, gpu-computing
- domain: deep-learning, machine-learning, computer-vision, gpu-computing
- platform: python, cpp
- tags: jit-compilation, meta-operators, pytorch-alternative, tsinghua, model-zoo, differentiable-rendering, linux, gpu, cuda

## Member repositories
- Jittor/jittor (main) score 67

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:07:50.095716+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 2026-08-30T07:24:40.461878+00:00, confidence not recorded.
  - readme: https://github.com/Jittor/jittor (fetched 2026-08-28T04:07:50.095716+00:00, sha 1559f9198909)
  - homepage: https://cg.cs.tsinghua.edu.cn/jittor/ (fetched 2026-08-29T09:37:42.686139+00:00, sha fe2a9dc730cb)
  - site_page: https://cg.cs.tsinghua.edu.cn/jittor/assets/docs/index.html (fetched 2026-08-29T09:37:42.695508+00:00, sha ceb1bbf73c3e)
  - site_page: https://cg.cs.tsinghua.edu.cn/jittor/about (fetched 2026-08-29T09:37:42.697479+00:00, sha f33143d18f09)
- Data as of 2026-08-30T08:39:29.467469+00:00.