|
|
--- |
|
|
dataset_info: |
|
|
features: |
|
|
- name: conversations |
|
|
list: |
|
|
- name: content |
|
|
dtype: string |
|
|
- name: role |
|
|
dtype: string |
|
|
- name: turn_identifier |
|
|
dtype: int64 |
|
|
- name: conversation_id |
|
|
dtype: string |
|
|
- name: source |
|
|
dtype: string |
|
|
splits: |
|
|
- name: train |
|
|
num_bytes: 137076991 |
|
|
num_examples: 37152 |
|
|
download_size: 74078361 |
|
|
dataset_size: 137076991 |
|
|
configs: |
|
|
- config_name: default |
|
|
data_files: |
|
|
- split: train |
|
|
path: data/train-* |
|
|
--- |
|
|
|
|
|
Regeneration of [shisa-ai/rewild-set](https://huggingface.co/datasets/shisa-ai/rewild-set) using DeepSeek-V3-0324 |
|
|
|
|
|
37152/180490 samples (20.58% of the original set) |
|
|
|
|
|
~13.5h on 10 x H100 nodes |
