---
dataset_info:
  features:
  - name: episode_index
    dtype: int64
  - name: task
    dtype: string
  - name: trajectory
    sequence:
      sequence: float64
  - name: prediction_raw
    dtype: string
  - name: num_trajectory_points
    dtype: int64
  - name: error
    dtype: string
  - name: source_dataset
    dtype: string
  - name: source_split
    dtype: string
  - name: observation_trace
    dtype: image
  splits:
  - name: train
    num_bytes: 1139281.0
    num_examples: 40
  download_size: 1117040
  dataset_size: 1139281.0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---