Update README.md
Browse files
README.md
CHANGED
|
@@ -5,4 +5,32 @@ tags:
|
|
| 5 |
- code llm
|
| 6 |
size_categories:
|
| 7 |
- n<1K
|
| 8 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 5 |
- code llm
|
| 6 |
size_categories:
|
| 7 |
- n<1K
|
| 8 |
+
---
|
| 9 |
+
|
| 10 |
+
<h1 style="text-align: center;">PythonSaga</h1>
|
| 11 |
+
This dataset follows the rules and diversity of template suggested in the paper "PythonSaga: Redefining the Benchmark to Evaluate Code Generating LLM" The goal is to make benchmarks better at assessing Code Generating Language Models (LLMs).
|
| 12 |
+
<br>
|
| 13 |
+
<br>
|
| 14 |
+
|
| 15 |
+
| **Model** | **Size** | **Pass@1** | **Pass@10** |
|
| 16 |
+
|-------------------------------|----------|------------|-------------|
|
| 17 |
+
| StarCoderBase | 7B | 0.0029 | 0.0149 |
|
| 18 |
+
| StarCoder2 | 7B | 0.0024 | 0.0217 |
|
| 19 |
+
| Code Llama | 7B | 0.0067 | 0.0472 |
|
| 20 |
+
| CodeQwen1.5-Chat | 7B | 0.0059 | 0.0497 |
|
| 21 |
+
| Nxcode-CQ-orpo | 7B | 0.0058 | 0.0523 |
|
| 22 |
+
| Mistral-Instruct-v0.1 | 7B | 0.0140 | 0.0552 |
|
| 23 |
+
| Code Llama Instruct | 7B | 0.0178 | 0.0744 |
|
| 24 |
+
| Deepseek Coder Instruct | 6.7B | 0.0137 | 0.0889 |
|
| 25 |
+
| Code Llama Python | 7B | 0.0240 | 0.0979 |
|
| 26 |
+
| Llama 3 | 8B | 0.0370 | 0.1125 |
|
| 27 |
+
| Phi-2 | 2.7B | 0.0302 | 0.1187 |
|
| 28 |
+
| OpenCodeInterpreter-DS | 6.7B | 0.0259 | 0.1206 |
|
| 29 |
+
| Deepseek Coder | 6.7B | 0.0343 | 0.1415 |
|
| 30 |
+
| Code Llama Python | 13B | 0.0405 | 0.1514 |
|
| 31 |
+
| GPT-3.5 | NA | 0.0724 | 0.2384 |
|
| 32 |
+
| GPT-4 | NA | 0.1243 | 0.3311 |
|
| 33 |
+
|
| 34 |
+
|
| 35 |
+
*Comparison between open and closed-source models on PythonSaga. We use the number of samples (n)
|
| 36 |
+
as 20 for both open and closed-source models.*
|