# baaivision/EVA

EVA Series: Visual Representation Fantasies from BAAI

Repository: https://github.com/baaivision/EVA
Canonical: https://ross.abutalabs.com/products/eva
Language: Python
License: MIT
License Family: permissive
Topics: foundation-models, representation-learning, vision-transformer
Last push: 2024-08-01T14:49:48+00:00

## Health v2 (maintenance only)
Score: 23/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 0, release rhythm 8, longevity 99
- inputs: {"age_days": 1388, "days_push": 762, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2691, forks 187 (observed 2026-08-28T04:07:10.962313+00:00)

## What it is
EVA is a family of large-scale vision foundation models from BAAI, including masked image models (EVA-01/02) and scaled CLIP models (EVA-CLIP up to 18B parameters). It provides pretrained weights and training code, distributed via Hugging Face, timm, and open_clip.

## Use cases
- extract visual features from images with a pretrained vision transformer
- fine-tune a large vision foundation model on my own image classification dataset
- use a large CLIP model for zero-shot image classification
- compute image and text embeddings for image search
- backbone for object detection or segmentation models

## When to avoid
- you need lightweight models for edge or CPU-only deployment
- you need a turnkey application rather than model weights and training code

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning
- domain: computer-vision, deep-learning, image-processing
- platform: python
- tags: vision-transformer, foundation-models, representation-learning, clip, masked-image-modeling, pretrained-models, gpu

## Member repositories
- baaivision/EVA (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:07:10.962313+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 2026-08-30T02:16:33.228269+00:00, confidence not recorded.
  - readme: https://github.com/baaivision/EVA (fetched 2026-08-28T04:07:10.962313+00:00, sha a89da83948bf)
- Data as of 2026-08-30T08:39:29.467469+00:00.