|
--- |
|
license: odc-by |
|
task_categories: |
|
- text-generation |
|
language: |
|
- en |
|
tags: |
|
- language-modeling |
|
- casual-lm |
|
- llm |
|
pretty_name: Dolma |
|
size_categories: |
|
- 100B<n<1T |
|
--- |
|
|
|
Tokenized (Llama 2) verison of [NousResearch/dolma-v1_7-30B](https://huggingface.co./datasets/NousResearch/dolma-v1_7-30B) as a [Nanotron](https://github.com/huggingface/nanotron) dataset split into 10 GB chunks. |
|
|
|
To download: |
|
```shell |
|
huggingface-cli download --repo-type dataset --local-dir dolma-v1_7-30B-tokenized-llama2-nanoset --local-dir-use-symlinks False NousResearch/dolma-v1_7-30B-tokenized-llama2-nanoset |
|
``` |
|
|
|
To recombine: |
|
```shell |
|
cat dolma-v1_7-30B-tokenized-llama2-nanoset_input_ids.npy.* > dolma-v1_7-30B-tokenized-llama2-nanoset_input_ids.npy |
|
rm -rf dolma-v1_7-305B-tokenized-llama3-nanoset |
|
``` |
|
|
|
Can also be used directly with numpy, for example |
|
```python |
|
import numpy as np |
|
|
|
dataset_buffer_mmap = np.memmap("dolma-v1_7-30B-tokenized-llama2-nanoset_input_ids.npy", |
|
mode="r", order="C", dtype=np.int32) |
|
dataset_buffer = memoryview(dataset_buffer_mmap) |
|
dataset_number_of_tokens = int(len(dataset_buffer)) |
|
``` |
