finnianx commited on
Commit
79a2ee5
·
verified ·
1 Parent(s): b86d22d

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +5 -3
README.md CHANGED
@@ -3,7 +3,9 @@ language:
3
  - en
4
  ---
5
  A tiny 55 million parameter trained on 1.3 billion tokens on a custom dataset mixture.
6
- HuggingFaceFW/fineweb-edu 50%
7
- epfml/FineWeb-HQ 30%
8
- HuggingFaceTB/cosmopedia (stories split) 20%
 
 
9
  The tokenizer is a basic bpe tokenizer that was trained on a smaller subset of 80_000 samples of this same data mixture with a vocab size of 8000.
 
3
  - en
4
  ---
5
  A tiny 55 million parameter trained on 1.3 billion tokens on a custom dataset mixture.
6
+ | Dataset | Weight |
7
+ |---|---|
8
+ | `HuggingFaceFW/fineweb-edu` | 50% |
9
+ | `epfml/FineWeb-HQ` | 30% |
10
+ | `HuggingFaceTB/cosmopedia` (stories split) | 20% |
11
  The tokenizer is a basic bpe tokenizer that was trained on a smaller subset of 80_000 samples of this same data mixture with a vocab size of 8000.