README.md

---
license: mit
dataset_info:
  features:
  - name: latent
    sequence:
      sequence:
        sequence: float32
  - name: noised_latents
    sequence:
      sequence:
        sequence: float32
  - name: noise
    sequence:
      sequence:
        sequence:
          sequence: float32
  - name: timesteps
    dtype: float64
  - name: label
    dtype: int64
  splits:
  - name: train
    num_bytes: 212160000
    num_examples: 60000
  download_size: 288948028
  dataset_size: 212160000
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---