Datasets:
Tasks:
Text Generation
Modalities:
Text
Formats:
csv
Languages:
English
Size:
1K - 10K
ArXiv:
Tags:
code generation
License:
Update README.md
Browse files
README.md
CHANGED
|
@@ -1,5 +1,11 @@
|
|
| 1 |
---
|
| 2 |
license: apache-2.0
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
---
|
| 4 |
|
| 5 |
# Codev-Bench
|
|
@@ -17,7 +23,7 @@ In detail, first, We extract unit test classes and functions from real GitHub re
|
|
| 17 |
|
| 18 |
We split the completion sub-scenes or capabilities that users may encounter while developing in an IDE into the following parts:
|
| 19 |
|
| 20 |
-
|
| 21 |
|
| 22 |
- **Scene1.1**. ✅ The context of the code block to be completed is fully complete.
|
| 23 |
|
|
@@ -25,15 +31,15 @@ We split the completion sub-scenes or capabilities that users may encounter whil
|
|
| 25 |
|
| 26 |
- **Scene1.3**. ✅ The context following the code block to be completed is completely empty.
|
| 27 |
|
| 28 |
-
|
| 29 |
|
| 30 |
- **Scene2.1**. ✅ Complete a portion of code within the code block.
|
| 31 |
|
| 32 |
- **Scene2.2**. ✅ The code block is already complete and should not have any code added.
|
| 33 |
|
| 34 |
-
|
| 35 |
|
| 36 |
-
|
| 37 |
|
| 38 |
|
| 39 |
## How To Use
|
|
@@ -188,4 +194,4 @@ myenv/bin/python src/evaluate.py --method print_scores --model codegemma_7b
|
|
| 188 |
|
| 189 |
We evaluate some popular general LLMs and code LLMs on the sub dataset **Scene1.1** of the CodevBench dataset. The results are as follows:
|
| 190 |
|
| 191 |
-
|
|
|
|
| 1 |
---
|
| 2 |
license: apache-2.0
|
| 3 |
+
task_categories:
|
| 4 |
+
- text-generation
|
| 5 |
+
language:
|
| 6 |
+
- en
|
| 7 |
+
tags:
|
| 8 |
+
- code
|
| 9 |
---
|
| 10 |
|
| 11 |
# Codev-Bench
|
|
|
|
| 23 |
|
| 24 |
We split the completion sub-scenes or capabilities that users may encounter while developing in an IDE into the following parts:
|
| 25 |
|
| 26 |
+
**Scene1**. ✅ Completing complete code blocks (including functions, conditional logic blocks, loop logic blocks, comments, ordinary statements, etc.).
|
| 27 |
|
| 28 |
- **Scene1.1**. ✅ The context of the code block to be completed is fully complete.
|
| 29 |
|
|
|
|
| 31 |
|
| 32 |
- **Scene1.3**. ✅ The context following the code block to be completed is completely empty.
|
| 33 |
|
| 34 |
+
**Scene2**. ✅ Completing a portion of code within a specific code block.
|
| 35 |
|
| 36 |
- **Scene2.1**. ✅ Complete a portion of code within the code block.
|
| 37 |
|
| 38 |
- **Scene2.2**. ✅ The code block is already complete and should not have any code added.
|
| 39 |
|
| 40 |
+
**Scene3**. 🔄 Completing code based on classes and functions defined in other files.
|
| 41 |
|
| 42 |
+
**Scene4**. 🔄 Completing code based on related and similar code within the project.
|
| 43 |
|
| 44 |
|
| 45 |
## How To Use
|
|
|
|
| 194 |
|
| 195 |
We evaluate some popular general LLMs and code LLMs on the sub dataset **Scene1.1** of the CodevBench dataset. The results are as follows:
|
| 196 |
|
| 197 |
+
|