This model is the MaskBit tokenizer with a vocabulary size of 16bits. It uses a downsampling factor of 16 and is trained on ImageNet for images of resolution 256.

You can find more details on the project page and in the paper.

Downloads last month: -; Downloads are not tracked for this model. How to track

Inference API

Unable to determine this model's library. Check the docs .

Dataset used to train markweber/maskbit_tokenizer_16bit

Collection including markweber/maskbit_tokenizer_16bit

maskbit

Collection

6 items • Updated 20 days ago

Evaluation results

rFID on ILSVRC/imagenet-1k
self-reported

1.290
InceptionScore on ILSVRC/imagenet-1k
self-reported

193.600
LPIPS on ILSVRC/imagenet-1k
self-reported

0.278
PSNR on ILSVRC/imagenet-1k
self-reported

21.800
SSIM on ILSVRC/imagenet-1k
self-reported

0.580
CodebookUsage on ILSVRC/imagenet-1k
self-reported

1.000

View on Papers With Code