# modelscope/modelscope

ModelScope: bring the notion of Model-as-a-Service to life.

Repository: https://github.com/modelscope/modelscope
Canonical: https://ross.abutalabs.com/products/modelscope
Homepage: https://www.modelscope.cn/
Language: Python
License: Apache-2.0
License Family: permissive
Topics: nlp, cv, speech, multi-modal, science, deep-learning, machine-learning, python
Last push: 2026-08-25T14:21:39+00:00

## Health v2 (maintenance only)
Score: 98/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 99, release rhythm 96, longevity 100
- inputs: {"age_days": 1500, "days_push": 8, "days_rel": 29, "gap_med": 14, "n_releases_24m": 46}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 9111, forks 962 (observed 2026-08-28T04:10:28.443865+00:00)

## What it is
ModelScope is a Python library and ecosystem built on the 'Model-as-a-Service' concept, providing unified APIs to download, run inference on, fine-tune, and evaluate machine learning models across CV, NLP, speech, multi-modal, and scientific computing domains. It aggregates state-of-the-art models from the AI community and streamlines their use in real-world applications with just a few lines of code.

## Use cases
- download and run pretrained models for inference
- fine-tune models on custom datasets
- evaluate machine learning models with a unified API
- find and use models for NLP, CV, speech, and multi-modal tasks
- integrate AI models into applications with minimal code
- contribute models to a shared model hub

## When to choose
- you want a unified API across many model domains like CV, NLP, and speech
- you want access to models hosted on the ModelScope hub, including China-hosted models
- you need simple few-line inference, fine-tuning, and evaluation workflows
- you prefer an alternative or complement to Hugging Face Transformers

## When to avoid
- you only need models available on Hugging Face and not on ModelScope
- you need a lightweight tool without the broader ModelScope ecosystem dependencies
- your workflow is tightly coupled to another ML framework's tooling

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, llm-inference, llm-training, nlp, image-processing, speech-recognition, sdk
- domain: machine-learning, deep-learning, large-language-models, computer-vision, speech-processing, artificial-intelligence
- platform: python, windows
- tags: model-hub, model-as-a-service, fine-tuning, model-evaluation, multi-modal, inference, huggingface-alternative, natural-language-processing, linux, macos, gpu

## Member repositories
- modelscope/modelscope (main) score 98

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:10:28.443865+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 2026-08-29T17:23:26.720983+00:00, confidence not recorded.
  - readme: https://github.com/modelscope/modelscope (fetched 2026-08-28T04:10:28.443865+00:00, sha b795d63a64fc)
  - homepage: https://www.modelscope.cn/ (fetched 2026-08-29T08:23:42.497797+00:00, sha b92e2c11ffac)
  - registry_pypi: https://pypi.org/pypi/modelscope/json (fetched 2026-08-29T08:23:42.500824+00:00, sha 57491340f892)
- Data as of 2026-08-30T08:39:29.467469+00:00.