README.md

metadata

license: apache-2.0
size_categories:
  - 100K<n<1M
pretty_name: evalhrf
configs:
  - config_name: ddpm-celebahq
    data_files:
      - split: train
        path: data/ddpm-celebahq/train/batch_*.zip
  - config_name: hrf-compressibility-window-baseline1
    data_files:
      - split: train
        path: data/hrf-compressibility-window-baseline-seed1/train/batch_*.zip
  - config_name: hrf-compressibility-window-baseline2
    data_files:
      - split: train
        path: data/hrf-compressibility-window-baseline-seed2/train/batch_*.zip
  - config_name: hrf-compressibility-window-baseline3
    data_files:
      - split: train
        path: data/hrf-compressibility-window-baseline-seed3/train/batch_*.zip
  - config_name: hrf-compressibility-window-early1
    data_files:
      - split: train
        path: data/hrf-compressibility-window-early-seed1/train/batch_*.zip
  - config_name: hrf-compressibility-window-early2
    data_files:
      - split: train
        path: data/hrf-compressibility-window-early-seed2/train/batch_*.zip
  - config_name: hrf-compressibility-window-early3
    data_files:
      - split: train
        path: data/hrf-compressibility-window-early-seed3/train/batch_*.zip
  - config_name: hrf-compressibility-window-later1
    data_files:
      - split: train
        path: data/hrf-compressibility-window-later-seed1/train/batch_*.zip
  - config_name: hrf-compressibility-window-later2
    data_files:
      - split: train
        path: data/hrf-compressibility-window-later-seed2/train/batch_*.zip
  - config_name: hrf-compressibility-window-later3
    data_files:
      - split: train
        path: data/hrf-compressibility-window-later-seed3/train/batch_*.zip
  - config_name: hrf-incompressibility-window-baseline1
    data_files:
      - split: train
        path: data/hrf-incompressibility-window-baseline-seed1/train/batch_*.zip
  - config_name: hrf-incompressibility-window-baseline2
    data_files:
      - split: train
        path: data/hrf-incompressibility-window-baseline-seed2/train/batch_*.zip
  - config_name: hrf-incompressibility-window-baseline3
    data_files:
      - split: train
        path: data/hrf-incompressibility-window-baseline-seed3/train/batch_*.zip
  - config_name: hrf-incompressibility-window-early1
    data_files:
      - split: train
        path: data/hrf-incompressibility-window-early-seed1/train/batch_*.zip
  - config_name: hrf-incompressibility-window-early2
    data_files:
      - split: train
        path: data/hrf-incompressibility-window-early-seed2/train/batch_*.zip
  - config_name: hrf-incompressibility-window-early3
    data_files:
      - split: train
        path: data/hrf-incompressibility-window-early-seed3/train/batch_*.zip
  - config_name: hrf-incompressibility-window-later1
    data_files:
      - split: train
        path: data/hrf-incompressibility-window-later-seed1/train/batch_*.zip
  - config_name: hrf-incompressibility-window-later2
    data_files:
      - split: train
        path: data/hrf-incompressibility-window-later-seed2/train/batch_*.zip
  - config_name: hrf-incompressibility-window-later3
    data_files:
      - split: train
        path: data/hrf-incompressibility-window-later-seed3/train/batch_*.zip
dataset_info:
  - config_name: ddpm-celebahq
    features:
      - name: image
        dtype: image
      - name: reward
        dtype: float
  - config_name: hrf-compressibility-window-baseline1
    features:
      - name: image
        dtype: image
      - name: reward
        dtype: float
  - config_name: hrf-compressibility-window-baseline2
    features:
      - name: image
        dtype: image
      - name: reward
        dtype: float
  - config_name: hrf-compressibility-window-baseline3
    features:
      - name: image
        dtype: image
      - name: reward
        dtype: float
  - config_name: hrf-compressibility-window-early1
    features:
      - name: image
        dtype: image
      - name: reward
        dtype: float
  - config_name: hrf-compressibility-window-early2
    features:
      - name: image
        dtype: image
      - name: reward
        dtype: float
  - config_name: hrf-compressibility-window-early3
    features:
      - name: image
        dtype: image
      - name: reward
        dtype: float
  - config_name: hrf-compressibility-window-later1
    features:
      - name: image
        dtype: image
      - name: reward
        dtype: float
  - config_name: hrf-compressibility-window-later2
    features:
      - name: image
        dtype: image
      - name: reward
        dtype: float
  - config_name: hrf-compressibility-window-later3
    features:
      - name: image
        dtype: image
      - name: reward
        dtype: float
  - config_name: hrf-incompressibility-window-baseline1
    features:
      - name: image
        dtype: image
      - name: reward
        dtype: float
  - config_name: hrf-incompressibility-window-baseline2
    features:
      - name: image
        dtype: image
      - name: reward
        dtype: float
  - config_name: hrf-incompressibility-window-baseline3
    features:
      - name: image
        dtype: image
      - name: reward
        dtype: float
  - config_name: hrf-incompressibility-window-early1
    features:
      - name: image
        dtype: image
      - name: reward
        dtype: float
  - config_name: hrf-incompressibility-window-early2
    features:
      - name: image
        dtype: image
      - name: reward
        dtype: float
  - config_name: hrf-incompressibility-window-early3
    features:
      - name: image
        dtype: image
      - name: reward
        dtype: float
  - config_name: hrf-incompressibility-window-later1
    features:
      - name: image
        dtype: image
      - name: reward
        dtype: float
  - config_name: hrf-incompressibility-window-later2
    features:
      - name: image
        dtype: image
      - name: reward
        dtype: float
  - config_name: hrf-incompressibility-window-later3
    features:
      - name: image
        dtype: image
      - name: reward
        dtype: float

Dataset Card for Eval Finetuning Diffusion Models with Reinforcement Learning

XYZ