Skip to content

EthenEthenEthen

LoginPlatformTry Ethen

Open Source Model Profile · facebook

deit-small-distilled-patch16-224

deit-small-distilled-patch16-224 is a distilled DeiT image classifier from Facebook trained on ImageNet-1k at 224x224. Its card documents distillation-token learning and a published 81.2% top-1 accuracy row.

Publisher
facebook
Task
image-classification
Model type
deit
License
apache-2.0
Library
transformers
Publication status
Accepted · not indexed

Model overview

deit-small-distilled-patch16-224 is published by Facebook as an image-classification model with DeiTForImageClassificationWithTeacher architecture and a deit model type. According to the model card, it is a small distilled Data-efficient Image Transformer pre-trained and fine-tuned with distillation on ImageNet-1k at 224x224. Captured metadata records an Apache-2.0 license and Transformers library compatibility.

Recorded capabilities

Distillation-token design

The card describes a distillation token that interacts with class and patch tokens through self-attention to learn from the teacher.

Documented inference preprocessing

The card specifies 256x256 resizing, 224x224 center-cropping, and ImageNet mean and standard-deviation normalization.

Published accuracy row

The card's evaluation table attributes 81.2% top-1 and 95.4% top-5 ImageNet accuracy to this distilled small variant.

Use cases in the source record

  • Image classification of photographs, such as COCO 2017 images, into one of the 1,000 ImageNet classes.
  • Accuracy comparisons across the DeiT family using the card's published top-1 and top-5 table.

Limitations and unknowns

  • No Safetensors parameter count was extracted from this record.
  • Accuracy figures are publisher-reported table values, not independently measured Ethen evaluations.
  • Provider state is historical snapshot data, not independently refreshed current availability.
  • The model card was written by the Hugging Face team, not the original releasing team.

Source and provenance

Source: facebook/deit-small-distilled-patch16-224

Captured: Unknown. Processed: 2026-09-07T19:34:44.394385+00:00.

Distilled Data-efficient Image Transformer (small-sized model) Distilled data-efficient Image Transformer (DeiT) model pre-trained and fine-tuned on ImageNet-1k (1 million images, 1,000 classes) at resolution 224x224. It was first introduced in the paper Training data-efficient image transformers & distillation through attention by Touvron et al. and first released in this repository . However, the weights were converted from the timm repository by Ross Wightman. Disclaimer: The team releasing DeiT did not write a model card for this model so this model card has been written by the Hugging Face team. Model description This model is…

F001F002F003F004F005F006F009F010F011F012F014F015F017F019