blob: 32daaa9292e9550a3cb1ec3a2575882c750e058d (
plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
|
## Quick Start
1. Download benchmark datasets:
```bash
cd OpenCodeEval/data
bash dataset.sh
```
2. Install dependencies:
```bash
pip install -e .
```
3. **Configure Evaluation Scripts**
- Replace placeholders in the evaluation scripts with the actual model name and path.
- Adjust any other necessary settings (e.g., evaluation parameters, output paths) to suit your requirements.
4. Execute the evaluation script for your desired benchmark. For example, to evaluate using the `test-humaneval-ckpt-list.sh` script:
Such as:
```bash
bash test-humaneval-ckpt-list.sh
```
> **Note**: Ensure that all configurations are correctly set before running the script to avoid errors.
|
