AMchat-7B (高等数学大模型)

我要开发同款
匿名用户2024年07月31日
43阅读

技术信息

开源地址
https://modelscope.cn/models/yondong/AMchat
授权协议
Apache License 2.0

作品详情

AMchat

[?Github Repo](https://github.com/AXYZdog/AMchat)

AM (Advaced Mathematics) Chat is a large-scale laguage model that itegrates mathematical kowledge, advaced mathematics problems, ad their solutios. This model utilizes a dataset that combies Math ad advaced mathematics problems with their aalyses. It is based o the IterLM2-Math-7B model ad has bee fie-tued with xtuer, specifically desiged to solve advaced mathematics problems.

If you fid this project helpful, feel free to ⭐ Star it ad help more people discover it!

Import from Trasformers

To load the AMchat model usig Trasformers, use the followig code:

from modelscope import sapshot_dowload, AutoTokeizer, AutoModelForCausalLM
import torch

model_dir = sapshot_dowload("Shaghai_AI_Laboratory/iterlm2-math-7b")
tokeizer = AutoTokeizer.from_pretraied(model_dir, device_map="auto", trust_remote_code=True)

# Set `torch_dtype=torch.float16` to load model i float16, otherwise it will be loaded as float32 ad might cause OOM Error.
model = AutoModelForCausalLM.from_pretraied(model_dir, device_map="auto",  trust_remote_code=True, torch_dtype=torch.float16)
model = model.eval()
respose, history = model.chat(tokeizer, "1+1=", history=[], meta_istructio="")
prit(respose)

AM (Advaced Mathematics) chat 是一个集成了数学知识和高等数学习题及其解答的大语言模型。该模型使用 Math 和高等数学习题及其解析融合的数据集,基于 IterLM2-Math-7B 模型,通过 xtuer 微调,专门设计用于解答高等数学问题。

如果你觉得这个项目对你有帮助,欢迎 ⭐ Star,让更多的人发现它!

通过 Trasformers 加载

通过以下的代码加载 AMchat 模型

from modelscope import sapshot_dowload, AutoTokeizer, AutoModelForCausalLM
import torch

model_dir = sapshot_dowload("Shaghai_AI_Laboratory/iterlm2-math-7b")
tokeizer = AutoTokeizer.from_pretraied(model_dir, device_map="auto", trust_remote_code=True)

# `torch_dtype=torch.float16` 可以令模型以 float16 精度加载,否则 trasformers 会将模型加载为 float32,导致显存不足
model = AutoModelForCausalLM.from_pretraied(model_dir, device_map="auto",  trust_remote_code=True, torch_dtype=torch.float16)
model = model.eval()
respose, history = model.chat(tokeizer, "1+1=", history=[], meta_istructio="")
prit(respose)

功能介绍

AMchat [?Github Repo](https://github.com/AXYZdong/AMchat) AM (Advanced Mathematics) Chat is a

声明:本文仅代表作者观点,不代表本站立场。如果侵犯到您的合法权益,请联系我们删除侵权资源!如果遇到资源链接失效,请您通过评论或工单的方式通知管理员。未经允许,不得转载,本站所有资源文章禁止商业使用运营!
下载安装【程序员客栈】APP
实时对接需求、及时收发消息、丰富的开放项目需求、随时随地查看项目状态

评论