Datasets:

CrowdMind
/

bash-smoltalk-small

pretty_name: Bash Instruct and Smol-SmolTalk
license: other
task_categories:
  - text-generation
language:
  - en
tags:
  - conversational
  - instruction-tuning
  - synthetic
configs:
  - config_name: default
    default: true
    data_files:
      - split: train
        path: data/train-*.parquet
      - split: test
        path: data/test-*.parquet
dataset_info:
  features:
    - name: messages
      list:
        - name: role
          dtype: string
        - name: content
          dtype: string
    - name: source_dataset
      dtype: string
    - name: source
      dtype: string
    - name: category
      dtype: string
    - name: utility
      dtype: string
    - name: variant_group
      dtype: string
  splits:
    - name: train
      num_examples: 514701
    - name: test
      num_examples: 24229

Bash Instruct and Smol-SmolTalk

This dataset combines Frost2o24/bash-instruct-III-55k and HuggingFaceTB/smol-smoltalk. The train split contains all rows from both source train splits and is globally shuffled with seed 42. The test split preserves the source test split from HuggingFaceTB/smol-smoltalk.

Split	Rows
train	514,701
test	24,229

The messages field is retained in chat format. source_dataset identifies the original dataset. Other columns preserve source metadata where available: source, category, utility, and variant_group; unavailable metadata is null.

from datasets import load_dataset

dataset = load_dataset("CrowdMind/bash-smoltalk-merged")
print(dataset)

Sources and licenses

Frost2o24/bash-instruct-III-55k is licensed under MIT.
HuggingFaceTB/smol-smoltalk is licensed under Apache-2.0.

This combined dataset contains material from both sources. Follow the applicable license and terms for each record, identified by source_dataset.