Datasets:
Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,51 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
task_categories:
|
| 4 |
+
- text-generation
|
| 5 |
+
- question-answering
|
| 6 |
+
language:
|
| 7 |
+
- en
|
| 8 |
+
tags:
|
| 9 |
+
- DeepSWE
|
| 10 |
+
- code
|
| 11 |
+
- coding
|
| 12 |
+
- programming
|
| 13 |
+
- SWE
|
| 14 |
+
- SWE-bench
|
| 15 |
+
- py
|
| 16 |
+
- js
|
| 17 |
+
- ts
|
| 18 |
+
- java
|
| 19 |
+
- cpp
|
| 20 |
+
- rust
|
| 21 |
+
- go
|
| 22 |
+
- reasoning
|
| 23 |
+
- SWE-smith
|
| 24 |
+
- gym
|
| 25 |
+
- agentic
|
| 26 |
+
size_categories:
|
| 27 |
+
- 10K<n<100K
|
| 28 |
+
---
|
| 29 |
+
# Dataset Description
|
| 30 |
+
|
| 31 |
+
This dataset is a heavily filtered version of all the SWE-bench/SWE-smith-lang datasets (expect php) merged together (originally 88,118 rows total), it aims to improve benchmark results on DeepSWE-style problems, benchmarks, and general coding skills
|
| 32 |
+
It is specifically filtered for rows with complex/long code problems
|
| 33 |
+
|
| 34 |
+
### Dataset Details
|
| 35 |
+
|
| 36 |
+
- **Curated by:** MoreThought
|
| 37 |
+
- **Funded by:** MoreThought
|
| 38 |
+
- **Shared by:** MoreThought
|
| 39 |
+
- **License:** MIT
|
| 40 |
+
|
| 41 |
+
### Dataset Sources
|
| 42 |
+
|
| 43 |
+
- **Repositorys:** All the SWE-bench/SWE-smith-lang datasets.
|
| 44 |
+
- **Paper:** https://huggingface.co/papers/2504.21798
|
| 45 |
+
|
| 46 |
+
## Uses
|
| 47 |
+
|
| 48 |
+
Improving benchmark results
|
| 49 |
+
Improving general coding capabilities
|
| 50 |
+
Training software engineering/coding agents
|
| 51 |
+
Improving long-context coding
|