File size: 3,742 Bytes
94612b8 687c9d9 94612b8 687c9d9 3a49dbf 687c9d9 5719e77 687c9d9 |
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 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 |
---
license: mit
datasets:
- TIGER-Lab/AceCode-89K
language:
- en
base_model:
- Qwen/Qwen2.5-Coder-7B-Instruct
tags:
- acecoder
- code
- Qwen
---
# 🂡 AceCode-89K
[Paper](https://arxiv.org/abs/2502.01718) |
[Github](https://github.com/TIGER-AI-Lab/AceCoder) |
[AceCode-89K](https://huggingface.co/datasets/TIGER-Lab/AceCode-89K) |
[AceCodePair-300K](https://huggingface.co/datasets/TIGER-Lab/AceCodePair-300K) |
[RM/RL Models](https://huggingface.co/collections/TIGER-Lab/acecoder-67a16011a6c7d65cad529eba)
We introduce AceCoder, the first work to propose a fully automated pipeline for synthesizing large-scale reliable tests used for the reward model training and reinforcement learning in the coding scenario. To do this, we curated the dataset AceCode-89K, where we start from a seed code dataset and prompt powerful LLMs to "imagine" proper test cases for the coding question and filter the noisy ones. We sample inferences from existing coder models and compute their pass rate as the reliable and verifiable rewards for both training the reward model and conducting the reinforcement learning for coder LLM.
## Note
- **This model is trained on the hard version of [TIGER-Lab/AceCode-89K](https://huggingface.co/datasets/TIGER-Lab/AceCode-89K) with about 22k examples, using the binary pass rate (rule based reward) as the reward**
<!-- - **This model official is trained on the hard version of [TIGER-Lab/AceCode-89K](https://huggingface.co/datasets/TIGER-Lab/AceCode-89K) with about 22k examples, using the [TIGER-Lab/AceCodeRM-7B](https://huggingface.co/TIGER-Lab/AceCodeRM-7B) as the reward** -->
- You can reproduce the hard version of [TIGER-Lab/AceCode-89K](https://huggingface.co/datasets/TIGER-Lab/AceCode-89K) using [script in our Github](#)
- The training takes 6 hours to finish on 8 x H100 GPUs in around 80 optimization steps.
- To reproduce the training, please refer to our [training script in the Github](#)
- To use the model, please refer to the codes in [Qwen/Qwen2.5-7B-Instruct](https://huggingface.co/Qwen/Qwen2.5-7B-Instruct)
- Training [wandb link](https://wandb.ai/dongfu/openrlhf_train_ppo/runs/5xqjy4uu)
## Usage
```python
from transformers import AutoModelForCausalLM, AutoTokenizer
model_name = "TIGER-Lab/AceCoder-Qwen2.5-Coder-7B-Ins-Rule"
model = AutoModelForCausalLM.from_pretrained(
model_name,
torch_dtype="auto",
device_map="auto"
)
tokenizer = AutoTokenizer.from_pretrained(model_name)
prompt = "Give me a short introduction to large language model."
messages = [
{"role": "system", "content": "You are Qwen, created by Alibaba Cloud. You are a helpful assistant."},
{"role": "user", "content": prompt}
]
text = tokenizer.apply_chat_template(
messages,
tokenize=False,
add_generation_prompt=True
)
model_inputs = tokenizer([text], return_tensors="pt").to(model.device)
generated_ids = model.generate(
**model_inputs,
max_new_tokens=512
)
generated_ids = [
output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
]
response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
```
## Performance
## Citation
```bibtex
@article{AceCoder,
title={AceCoder: Acing Coder RL via Automated Test-Case Synthesis},
author={Zeng, Huaye and Jiang, Dongfu and Wang, Haozhe and Nie, Ping and Chen, Xiaotong and Chen, Wenhu},
journal={ArXiv},
year={2025},
volume={abs/2207.01780}
}
``` |