jtatman's picture
format incorrect
8c2cb51
metadata
license: mit
dataset_info:
  features:
    - name: instruction
      dtype: string
    - name: context
      dtype: string
    - name: response
      dtype: string
    - name: category
      dtype: string
    - name: __index_level_0__
      dtype: int64
  splits:
    - name: train
    - num_bytes: 3853338
    - num_examples: 3663
    - download_size: 2447514
    - dataset_size: 3853338

this is a parse down of the esoteric categories in dolly 15k dataset the size is intentional for processing here on the hub

.::modification of the databricks 15k dataset for on hub processing::.