Eval Results
markweber's picture
Update README.md
7a2812b verified
metadata
license: apache-2.0
datasets:
  - ILSVRC/imagenet-1k
model-index:
  - name: MaskBit-Generator-14bits
    results:
      - task:
          type: image-generation
        dataset:
          name: ILSVRC/imagenet-1k
          type: ILSVRC/imagenet-1k
        metrics:
          - name: FID
            type: FID
            value: 1.52
          - name: InceptionScore
            type: InceptionScore
            value: 328.6

This model is the MaskBit generator with a vocabulary size of 14 bits. It is trained on ImageNet for images of resolution 256 and evaluated with 256 sampling steps.

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