数学问题求解模型
这是一个基于tungdqzenai/Grok_HumanLike_Llama-3.2-3B-Instruct 微调的数学问题求解模型。
模型描述
- 基础模型: tungdqzenai/Grok_HumanLike_Llama-3.2-3B-Instruct
- 微调任务: 数学推理和问题求解
- 参数量: 3B
- 语言: 中文、英文
使用方法
方法一:使用 transformers 库
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch
# 加载模型和分词器
model_name = "zhman/llama-SFT-GRPO"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(
model_name,
torch_dtype=torch.bfloat16,
device_map="auto"
)
# 推理示例
def solve_math_problem(question):
prompt = f"问题:{question}\n答案:"
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
outputs = model.generate(
**inputs,
max_length=512,
temperature=0.7,
top_p=0.9,
do_sample=True
)
answer = tokenizer.decode(outputs[0], skip_special_tokens=True)
return answer
# 测试
result = solve_math_problem("2+2等于多少?")
print(result)
方法二:使用 HuggingFace Inference API
import requests
API_URL = "https://api-inference.huggingface.co/models/zhman/llama-SFT-GRPO"
headers = {"Authorization": "Bearer YOUR_HF_TOKEN"}
def query(payload):
response = requests.post(API_URL, headers=headers, json=payload)
return response.json()
output = query({
"inputs": "问题:3×5等于多少?",
"parameters": {"max_length": 200, "temperature": 0.7}
})
print(output)
训练详情
- 训练框架: Unsloth
- 基础模型: tungdqzenai/Grok_HumanLike_Llama-3.2-3B-Instruct
- 数据类型: bfloat16
- 最大序列长度: 131072
限制与注意事项
- 模型专注于数学问题求解,其他领域的表现可能有限
- 推理时建议使用适当的 temperature 参数以平衡创造性和准确性
引用
如果您使用了本模型,请引用:
@misc{llama-math-tuned,
author = {Your Name},
title = {Llama Math Fine-tuned Model},
year = {2026},
publisher = {HuggingFace},
url = {https://huggingface.co/zhman/llama-SFT-GRPO}
}
- Downloads last month
- 2
Inference Providers
NEW
This model isn't deployed by any Inference Provider.
🙋
Ask for provider support
Model tree for zhman/llama-SFT-GRPO
Base model
meta-llama/Llama-3.2-3B-Instruct
Finetuned
unsloth/Llama-3.2-3B-Instruct