vit_mae¶


Model type	`vit_mae`
Class	`ViTModel`
Task	`image-classification`
Source	`models/vit.py`

Description¶

Vision Transformer for image feature extraction.

Pre-norm encoder with patch embeddings, CLS token, and learned position embeddings. Output is the last hidden state including the CLS token at position 0.

Usage¶

mobius build --model <MODEL_ID> output_dir/

from mobius import build

model = build("<MODEL_ID>")