mirror of
https://github.com/deepseek-ai/DeepSeek-Coder.git
synced 2025-04-20 02:29:09 -04:00
86 lines
4.5 KiB
Markdown
86 lines
4.5 KiB
Markdown
## 1. Introduction
|
|
|
|
We provide a test script to evaluate the capability of the **deepseek-coder** model to solve mathematical problems using external tools (Python interpreter). We evaluate it using the [PAL](https://arxiv.org/pdf/2211.10435.pdf) method on seven datasets: **GSM8k, MATH, GSM-Hard, SVAMP, TabMWP, ASDiv, and MAWPS**.
|
|
|
|
|
|
|
|
## 2. Setup
|
|
|
|
```
|
|
pip install sympy==1.12 pebble timeout-decorator transformers
|
|
```
|
|
|
|
|
|
|
|
## 3. Evaluation
|
|
|
|
We provide an example of testing the **deepseek-coder-1b-python** model on the **gsm8k** dataset using **8** GPUs. If you wish to use a different model or dataset, you can modify it as needed.
|
|
|
|
```bash
|
|
MODEL_NAME_OR_PATH=deepseek/deepseek-coder-1b-python
|
|
DATA=gsm8k # 'math' 'gsm8k' 'gsm-hard' 'svamp' 'tabmwp' 'asdiv' 'mawps'
|
|
MODEL_DIR_NAME=${MODEL_NAME_OR_PATH##*/}
|
|
GPU_NUM=8
|
|
for rank in {0..7}; do
|
|
CUDA_VISIBLE_DEVICES=$rank nohup python run.py \
|
|
--data_name ${DATA} \
|
|
--model_name_or_path ${MODEL_NAME_OR_PATH} \
|
|
--batch_size 16 \
|
|
--do_inference \
|
|
--rank $rank \
|
|
--world_size $GPU_NUM 2>&1 &
|
|
done
|
|
|
|
# Wait for all processes to finish
|
|
wait
|
|
echo "All processes completed."
|
|
python run.py --do_eval --data_name ${DATA} --model_name_or_path ${MODEL_NAME_OR_PATH} --world_size $GPU_NUM | tee outputs/${MODEL_DIR_NAME}/${DATA}/result.out
|
|
```
|
|
|
|
|
|
|
|
## 4. Experimental Results
|
|
|
|
We report experimental results here for mathematical reasoning tasks by using python program. For all open-source models, we utilize this repository and test with the same prompt. We set the maximum input length to **2048** and the maximum output length to **512**, and employ the **greedy search strategy**.
|
|
|
|
|
|
|
|
#### (1) Multilingual Base Models
|
|
|
|
| Model | Size | GSM8k | MATH | GSM-Hard | SVAMP | TabMWP | ASDiv | MAWPS | Avg |
|
|
| -------------- | ---- | ----- | ----- | -------- | ----- | ------ | ----- | ----- | ----- |
|
|
| CodeShell | 7B | 17.0% | 9.1% | 18.2% | 45.6% | 29.6% | 46.6% | 56.8% | 31.8% |
|
|
| CodeGeex-2 | 7B | 23.6% | 9.6% | 22.4% | 48.0% | 47.2% | 46.9% | 66.0% | 37.7% |
|
|
| StarCoder-Base | 16B | 27.3% | 11.5% | 24.2% | 44.0% | 45.6% | 54.9% | 73.4% | 40.1% |
|
|
| CodeLLama-Base | 7B | 36.4% | 12.3% | 29.7% | 57.6% | 58.4% | 59.6% | 82.6% | 48.0% |
|
|
| CodeLLama-Base | 13B | 44.2% | 15.5% | 42.4% | 65.6% | 61.6% | 65.3% | 85.3% | 54.3% |
|
|
| CodeLLama-Base | 34B | 58.2% | 22.1% | 55.2% | 77.2% | 69.6% | 70.0% | 92.8% | 63.6% |
|
|
| | | | | | | | | | |
|
|
| OraCoder-Base | 1B | 17.0% | 13.4% | 13.3% | 39.2% | 42.4% | 44.8% | 66.0% | 33.7% |
|
|
| OraCoder-Base | 7B | 46.0% | 20.6% | 40.0% | 67.2% | 71.2% | 67.1% | 89.1% | 57.3% |
|
|
| OraCoder-Base | 33B | - | - | - | - | - | - | - | - |
|
|
|
|
#### (2) Python Base Models
|
|
|
|
| Model | Size | GSM8k | MATH | GSM-Hard | SVAMP | TabMWP | ASDiv | MAWPS | Avg |
|
|
| -------------- | ---- | ----- | ----- | -------- | ----- | ------ | ----- | ----- | ----- |
|
|
| StarCoder | 16B | 31.5% | 13.8% | 26.7% | 48.8% | 47.2% | 54.9% | 76.1% | 42.7% |
|
|
| CodeLLama-Python | 7B | 35.2% | 14.7% | 34.5% | 70.4% | 55.2% | 62.1% | 84.2% | 50.9% |
|
|
| CodeLLama-Python | 13B | 44.8% | 17.4% | 45.5% | 65.6% | 60.8% | 69.0% | 89.6% | 56.1% |
|
|
| CodeLLama-Python | 34B | 57.6% | 21.1% | 54.5% | 76.8% | 66.8% | 69.5% | 94.2% | 62.9% |
|
|
| | | | | | | | | | |
|
|
| OraCoder-Python | 1B | 17.6% | 15.0% | 18.2% | 40.0% | 38.4% | 49.5% | 64.1% | 34.7% |
|
|
| OraCoder-Python | 7B | 50.3% | 24.3% | 43.0% | 71.2% | 73.6% | 69.7% | 88.0% | 60.0% |
|
|
| OraCoder-Python | 33B | - | - | - | - | - | - | - | - |
|
|
|
|
#### (3) Instruction-Tuned Models
|
|
| Model | Size | GSM8k | MATH | GSM-Hard | SVAMP | TabMWP | ASDiv | MAWPS | Avg |
|
|
| -------------- | ---- | ----- | ----- | -------- | ----- | ------ | ----- | ----- | ----- |
|
|
| ChatGPT | - | 78.6% | 38.7% | 67.6% | 77.8% | 79.9% | 81.0% | 89.4% | 73.3% |
|
|
| GPT-4 | - | 94.2% | 51.8% | 77.6% | 94.8% | 95.9% | 92.6% | 97.7% | 86.4% |
|
|
| | | | | | | | | | |
|
|
| OraCoder-Chat | 1B | - | - | - | - | - | - | - | - |
|
|
| OraCoder-Chat | 7B | - | - | - | - | - | - | - | - |
|
|
| OraCoder-Chat | 33B | - | - | - | - | - | - | - | - |
|
|
|