---
dataset_info:
  features:
  - name: qid
    dtype: string
  - name: question
    dtype: string
  - name: latent_states
    list:
      list: float32
  - name: answer_digits
    list: int32
  - name: num_latents
    dtype: int32
  - name: k_max
    dtype: int32
  - name: K_star
    dtype: int32
  splits:
  - name: train
    num_bytes: 1252932561
    num_examples: 54000
  download_size: 506043818
  dataset_size: 1252932561
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---