Model Description

This model is a fine-tuned variant of Qwen2.5-Coder-7B-Instruct, optimized for code-related tasks by training on a specialized code dataset. Fine-tuning was performed using the Hugging Face Transformers library, Low-Rank Adaptation(LoRA) and Parameter-Efficient Fine-Tuning (PEFT) techniques,.

  • Intended Use: The provided fine-tuned weights are ready for immediate use, making this model particularly suitable for developers and researchers working on code comprehension, generation, and related applications.
  • Performance: Users can anticipate enhanced performance on coding-specific tasks compared to the base model. The actual improvement varies based on the task and input data specifics.

Use

The base model Qwen2.5-Coder-7B-Instruct needs to be downloaded and loaded in advance, then provide the following information: "## Problem Description: {} ## Test Cases: {} ## Error Code: {}"

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch
from peft import PeftModel

model_path = 'Qwen/Qwen2.5-Coder-7B-Instruct'
lora_path = 'Code-AiHelper'  

tokenizer = AutoTokenizer.from_pretrained(model_path)
model = AutoModelForCausalLM.from_pretrained(model_path, device_map="auto",torch_dtype=torch.bfloat16)
model = PeftModel.from_pretrained(model, model_id=lora_path)

system_prompt = '你是一位经验丰富的Python编程专家和技术顾问,擅长分析Python题目和学生编写的代码。你的任务是理解题目要求和测试样例,分析学生代码,找出潜在的语法或逻辑错误,提供具体的错误位置和修复建议,并用专业且易懂的方式帮助学生改进代码。请以markdown格式返回你的答案。'
user_prompt = '''## 题目描述:{} ## 测试样例:{} ## 错误代码:{}'''
messages = [
    {"role": "system", "content": system_prompt},
    {"role": "user", "content": user_prompt}
]

text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
model_inputs = tokenizer([text], return_tensors="pt").to('cuda')
generated_ids = model.generate(
    model_inputs.input_ids,
    max_new_tokens=1024
)
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]

print(response)
Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model is not currently available via any of the supported Inference Providers.
The model cannot be deployed to the HF Inference API: The model has no library tag.

Model tree for monidew/Code-AiHelper

Base model

Qwen/Qwen2.5-7B
Finetuned
(45)
this model