Commit
•
c26986d
1
Parent(s):
ea47812
Update README.md
Browse files
README.md
CHANGED
|
@@ -24,3 +24,13 @@ configs:
|
|
| 24 |
- split: valid
|
| 25 |
path: data/valid-*
|
| 26 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 24 |
- split: valid
|
| 25 |
path: data/valid-*
|
| 26 |
---
|
| 27 |
+
|
| 28 |
+
# Pretraining dataset
|
| 29 |
+
This dataset is composed of 1. dolma 22B, 2. fineweb-edu 21B, and 3. starcoder 2B with common programming language (C, CPP, JAVA, PYTHON, JSON)
|
| 30 |
+
|
| 31 |
+
## Token count
|
| 32 |
+
|
| 33 |
+
| split | GPT2 token count | Llama3 token count |
|
| 34 |
+
| ---- | ---- | ---- |
|
| 35 |
+
| train | 45.25B | 41.80B |
|
| 36 |
+
| valid | 45.04M | 41.64M
|