# apple/ml-mobileclip

This repository contains the official implementation of the research papers, "MobileCLIP" CVPR 2024 and "MobileCLIP2" TMLR August 2025

Repository: https://github.com/apple/ml-mobileclip
Canonical: https://ross.abutalabs.com/products/ml-mobileclip
Language: Python
License: NOASSERTION
License Family: other
Last push: 2026-04-15T00:55:23+00:00

## Health v2 (maintenance only)
Score: 60/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 77, release rhythm 35, longevity 65
- inputs: {"age_days": 916, "days_push": 141, "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 1627, forks 130 (observed 2026-08-28T04:05:12.997656+00:00)

## What it is
Official implementation of Apple's MobileCLIP and MobileCLIP2, fast image-text contrastive (CLIP-style) models optimized for mobile latency. It includes inference, training, and evaluation code, pretrained checkpoints on HuggingFace, and an iOS demo app for real-time zero-shot image classification.

## Use cases
- run zero-shot image classification on mobile devices
- compute image and text embeddings with a fast CLIP model
- train a lightweight CLIP model with multi-modal reinforced training
- evaluate image-text models on ImageNet zero-shot benchmarks
- deploy CLIP models on iPhone with low latency
- compare MobileCLIP accuracy against OpenAI CLIP and SigLIP

## When to choose
- you need CLIP-style embeddings or zero-shot classification with low latency on mobile hardware
- you want small, efficient multimodal models with pretrained checkpoints
- you want to reproduce or extend MobileCLIP/MobileCLIP2 training

## When to avoid
- you need the largest, highest-accuracy CLIP models regardless of latency
- you need non-PyTorch or non-Apple-ecosystem deployment support out of the box
- you need a general-purpose vision-language framework rather than specific model implementations

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, image-processing, nlp
- domain: machine-learning, computer-vision, mobile-development
- platform: python
- tags: clip, zero-shot-classification, image-text-models, mobile-inference, multimodal, apple, natural-language-processing, ios, gpu

## Member repositories
- apple/ml-mobileclip (main) score 60

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:12.997656+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 2026-08-30T03:48:23.578333+00:00, confidence not recorded.
  - readme: https://github.com/apple/ml-mobileclip (fetched 2026-08-28T04:05:12.997656+00:00, sha 6e1c643b34e6)
- Data as of 2026-08-30T08:39:29.467469+00:00.