# google-ai-edge/litert-torch

Support PyTorch model conversion with LiteRT.

Repository: https://github.com/google-ai-edge/litert-torch
Canonical: https://ross.abutalabs.com/products/litert-torch
Language: Jupyter Notebook
License: Apache-2.0
License Family: permissive
Last push: 2026-09-02T22:00:08+00:00

## Health v2 (maintenance only)
Score: 87/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 100, release rhythm 87, longevity 60
- inputs: {"age_days": 842, "days_push": 0, "days_rel": 9, "gap_med": 55.0, "n_releases_24m": 9}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1091, forks 187 (observed 2026-09-03T02:15:12.376213+00:00)

## What it is
LiteRT Torch is a Python library that converts PyTorch models into .tflite format for running with LiteRT on Android, iOS, and IoT devices. It builds on torch.export() with broad CPU coverage and initial GPU/NPU support, including a Generative API for LLMs and transformer models.

## Use cases
- convert pytorch model to tflite
- run pytorch models on android on-device
- deploy llm on mobile device
- quantize transformer model for edge inference
- convert resnet to tflite flatbuffer
- run machine learning models on iot devices

## When to choose
- you have PyTorch models and need to run them on-device with LiteRT
- you want to deploy LLMs or transformer models on mobile with quantization
- you need broad Core ATen operator coverage for model conversion

## When to avoid
- you target TensorFlow/Keras models rather than PyTorch
- you need server-side or cloud inference rather than on-device
- you require stable production support for the Generative API, which is Alpha

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, llm-inference, compiler, serialization
- domain: machine-learning, deep-learning, large-language-models, mobile-development, embedded-systems
- platform: python, iot
- tags: pytorch, tflite, litert, on-device-ai, model-conversion, quantization, edge-ai, android, ios, linux

## Member repositories
- google-ai-edge/litert-torch (main) score 87

## Provenance
- Observed fields: from GitHub, fetched 2026-09-03T02:15:12.376213+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 2026-08-30T06:50:02.020381+00:00, confidence not recorded.
  - readme: https://github.com/google-ai-edge/litert-torch (fetched 2026-09-03T02:15:12.376213+00:00, sha 8ca77432fa7f)
  - registry_pypi: https://pypi.org/pypi/litert-torch/json (fetched 2026-08-29T12:53:03.747381+00:00, sha 66df5c41e54f)
- Data as of 2026-08-30T08:39:29.467469+00:00.