README.md

---
dataset_info:
  features:
  - name: env_name
    dtype: string
  - name: episode
    dtype: int64
  - name: step
    dtype: int64
  - name: images
    dtype: image
  - name: action_id
    dtype: int64
  - name: agent_pos
    list: int64
  - name: agent_dir
    dtype: int64
  - name: texts
    list:
    - name: user
      dtype: string
    - name: assistant
      dtype: string
  splits:
  - name: train
    num_bytes: 1907553
    num_examples: 967
  - name: train_clean
    num_bytes: 59121
    num_examples: 30
  - name: test
    num_bytes: 39453
    num_examples: 20
  download_size: 1996020
  dataset_size: 2006127
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: train_clean
    path: data/train_clean-*
  - split: test
    path: data/test-*
---