# microsoft/Cream

This is a collection of our NAS and Vision Transformer work.

Repository: https://github.com/microsoft/Cream
Canonical: https://ross.abutalabs.com/products/cream
Language: Python
License: MIT
License Family: permissive
Topics: nas, automl, vision-transformer, rpe, vit-compression, efficiency, knowledge-distillation
Last push: 2024-07-25T10:29:15+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 100
- inputs: {"age_days": 2151, "days_push": 769, "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 1839, forks 244 (observed 2026-08-28T04:05:42.891196+00:00)

## What it is
A collection of Microsoft research projects on neural architecture search (NAS) and efficient Vision Transformers, including AutoFormer, TinyViT, EfficientViT, MiniViT, TinyCLIP, and iRPE. It provides reference implementations, pretrained models, and training code for architecture search and ViT compression techniques.

## Use cases
- search for efficient vision transformer architectures
- compress vision transformers with knowledge distillation
- train small ViT models with distillation pretraining
- improve relative position encoding in vision transformers
- distill CLIP models into tiny multimodal encoders
- run one-shot neural architecture search for image models

## When to choose
- you need reference implementations of published NAS or efficient ViT papers
- you want pretrained tiny/efficient vision transformer checkpoints
- you are researching architecture search or ViT compression

## When to avoid
- you need a production-ready model training framework
- you want a maintained general-purpose deep learning library
- you need non-vision or non-transformer NAS tasks

## Facets
- artifact type: library
- maturity: maintenance
- function: machine-learning, deep-learning
- domain: machine-learning, computer-vision, deep-learning
- platform: python
- tags: nas, vision-transformer, automl, knowledge-distillation, model-compression, research-code, gpu

## Member repositories
- microsoft/Cream (main) score 23

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:05:42.891196+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 2026-08-30T03:18:14.866094+00:00, confidence not recorded.
  - readme: https://github.com/microsoft/Cream (fetched 2026-08-28T04:05:42.891196+00:00, sha d2d075ed93d9)
- Data as of 2026-08-30T08:39:29.467469+00:00.