Datasets:
Update README.md
Browse files
README.md
CHANGED
|
@@ -21,3 +21,43 @@ configs:
|
|
| 21 |
- split: train
|
| 22 |
path: data/train-*
|
| 23 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 21 |
- split: train
|
| 22 |
path: data/train-*
|
| 23 |
---
|
| 24 |
+
---
|
| 25 |
+
pretty_name: github-java-corpus
|
| 26 |
+
language:
|
| 27 |
+
- code
|
| 28 |
+
license: other
|
| 29 |
+
task_categories:
|
| 30 |
+
- text-generation
|
| 31 |
+
tags:
|
| 32 |
+
- text
|
| 33 |
+
- code
|
| 34 |
+
- java
|
| 35 |
+
---
|
| 36 |
+
|
| 37 |
+
# github-java-2012
|
| 38 |
+
|
| 39 |
+
## Summary
|
| 40 |
+
This dataset contains Java source-code text samples prepared for pretraining.
|
| 41 |
+
|
| 42 |
+
## Repository
|
| 43 |
+
`TheFinAI/github-java-corpus`
|
| 44 |
+
|
| 45 |
+
## Required Columns
|
| 46 |
+
- **Source**: dataset name
|
| 47 |
+
- **Date**: year
|
| 48 |
+
- **Text**: the pure text of each sample
|
| 49 |
+
- **Token_count**: the token count computed with tiktoken
|
| 50 |
+
|
| 51 |
+
## Schema
|
| 52 |
+
- `Source` (`string`)
|
| 53 |
+
- `Date` (`int32`)
|
| 54 |
+
- `Text` (`string`)
|
| 55 |
+
- `Token_count` (`int32`)
|
| 56 |
+
|
| 57 |
+
## Construction
|
| 58 |
+
The dataset was built from streamed archive processing into file-level text samples.
|
| 59 |
+
|
| 60 |
+
## Notes
|
| 61 |
+
- Intended for pretraining
|
| 62 |
+
- One row corresponds to one source file
|
| 63 |
+
- `Date` is the assigned year label for this source
|