# mlfoundations/open_clip

An open source implementation of CLIP.

Repository: https://github.com/mlfoundations/open_clip
Canonical: https://ross.abutalabs.com/products/open_clip
Language: Python
License: NOASSERTION
License Family: other
Topics: deep-learning, pytorch, computer-vision, language-model, multi-modal-learning, contrastive-loss, zero-shot-classification, pretrained-models
Last push: 2026-08-23T20:06:40+00:00

## Health v2 (maintenance only)
Score: 86/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 99, release rhythm 60, longevity 100
- inputs: {"age_days": 1862, "days_push": 10, "days_rel": 188, "gap_med": 43.5, "n_releases_24m": 11}
- flags: no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 14095, forks 1303 (observed 2026-08-28T04:11:05.473837+00:00)

## What it is
OpenCLIP is an open-source PyTorch implementation of CLIP and related multimodal contrastive models, with many pretrained image/text checkpoints. It supports both inference with pretrained models and large-scale contrastive training including newer architectures like CoCa, MaMMUT, and NaFlex variants.

## Use cases
- compute image and text embeddings with pretrained CLIP models
- zero-shot image classification from natural language prompts
- train a custom CLIP model on my own image-text dataset
- build image search by comparing text queries to image embeddings
- fine-tune contrastive image-text models
- extract features from images for downstream tasks

## When to choose
- you need CLIP-style multimodal embeddings in PyTorch
- you want zero-shot classification without training a custom classifier
- you need to pretrain or fine-tune contrastive image-text models at scale
- you want access to many open pretrained checkpoints (OpenAI, LAION, DataComp)

## When to avoid
- you need a non-PyTorch framework like TensorFlow or JAX
- you only need a hosted inference API without local GPU resources
- your task is single-modality (pure vision or pure NLP) with no cross-modal needs

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, computer-vision, nlp, llm-training
- domain: deep-learning, computer-vision, machine-learning
- platform: python
- tags: clip, contrastive-learning, pytorch, multimodal, zero-shot-classification, pretrained-models, image-text, natural-language-processing, gpu

## Member repositories
- mlfoundations/open_clip (main) score 86

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:11:05.473837+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 2026-08-29T17:12:49.253080+00:00, confidence not recorded.
  - readme: https://github.com/mlfoundations/open_clip (fetched 2026-08-28T04:11:05.473837+00:00, sha 9b14e426bd1e)
- Data as of 2026-08-30T08:39:29.467469+00:00.