File size: 363 Bytes
9903389 f08f8fa |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 |
---
license: mit
datasets:
- NobodyExistsOnTheInternet/GiftedConvoBeforeEcons
---
Trained on over 20k instruct generated all by gpt-4 or humans
Dataset features:
1000 long evolved conversations based off LIMA
Subsection of correct PRM800k data
Subsection of CamelAI's Physics and Chemistry data
The model is trained with Qlora as well as Axolotl. |