olmo-hybrid pretraining data are not available
Open
- Dominant language
- Python
- Stars
- 1.5k
- Forks
- 315
- Avg merge
- 1d 9h
- Merged PRs (30d)
- 11
Description
Thanks to the OLMo team for excellent open-source work. I tried to manually download the training data for OLMo-Hybrid, but the link seems to be broken. Could you please let me know what the correct link is? e.g. https://olmo-data.org/preprocessed/dolma2-0625/v0.1/allenai/dolma2-tokenizer/all-dressed-snazzy2/adult_content/000000.npy
```sh
wget https://olmo-data.org/preprocessed/dolma2-0625/v0.1/allenai/dolma2-tokenizer/all-dressed-snazzy2/adult_content/000000.npy
```
or this should be `OLMo_mix_0925_official` ?
Contributor guide
Assessment
This issue has not been assessed yet.