---
dataset_info:
  features:
  - name: text
    dtype: string
  - name: meta
    struct:
    - name: redpajama_set_name
      dtype: string
  splits:
  - name: train
    num_bytes: 0
    num_examples: 590394625
  - name: validation
    num_bytes: 0
    num_examples: 502556
  - name: test
    num_bytes: 0
    num_examples: 502268
  download_size: 1507040721942
  dataset_size: 0
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: validation
    path: data/validation-*
  - split: test
    path: data/test-*
---
As datasets puts limits on the number of calls to huggingface, downloading SlimPajama-627B is problematic as it's composed of a ton of small files. 
I have reuploaded it here as larger chunks to easily download the dataset without having to do anything hacky.

The original dataset can be found here [https://huggingface.co/datasets/cerebras/SlimPajama-627B](https://huggingface.co/datasets/cerebras/SlimPajama-627B)