D-X-Y / AutoDL-Projects

Automated deep learning algorithms implemented in PyTorch.
MIT License
1.57k stars 284 forks source link

How to run #101

Closed ZYF12138-zyf closed 3 years ago

ZYF12138-zyf commented 3 years ago

Man, how to run it? Help! The teacher found your code on GitHub as our final evaluation. The teacher was irresponsible and didn't teach us python. Help me!

D-X-Y commented 3 years ago

This repo contains multiple NAS algorithms. Which one do you prefer to running? If you have no preference, I would recommend NATS-Bench. In this case, please have a look at https://github.com/D-X-Y/AutoDL-Projects/tree/main/exps/NATS-algos and our paper (https://arxiv.org/abs/2009.00437).

PS: This issue is identical to #100

ZYF12138-zyf commented 3 years ago

Could you please tell me which one is easier to operate? Could you please recommend one or two algorithms that are easier to understand?

------------------ 原始邮件 ------------------ 发件人: @.>; 发送时间: 2021年5月6日(星期四) 晚上8:42 收件人: @.>; 抄送: @.>; @.>; 主题: Re: [D-X-Y/AutoDL-Projects] How to run (#101)

这个repo包含多个NAS算法。比起跑步,你更喜欢哪一种?如果你没有偏好,我会推荐NATS-Bench。在这种情况下,请您看看https://github.com/D-X-Y/AutoDL-Projects/tree/main/exps/NATS-algos而我们的论文(https://arxiv.org/abs/2009.00437).

— 您正在接收此消息,因为您编写了线程。 直接回复此邮件,在GitHub上查看,或取消订阅.

D-X-Y commented 3 years ago

NATS-Bench is easier to run. Commands are listed in the comments in each file under this folder: https://github.com/D-X-Y/AutoDL-Projects/tree/main/exps/NATS-algos

ZYF12138-zyf commented 3 years ago

Thank you, Sir, I am not engaged in this industry after graduation, but your kindness today, let me have the idea of engaged in this field, thank you, I wish you a happy life and a successful career!

------------------ 原始邮件 ------------------ 发件人: @.>; 发送时间: 2021年5月6日(星期四) 晚上8:58 收件人: @.>; 抄送: @.>; @.>; 主题: Re: [D-X-Y/AutoDL-Projects] How to run (#101)

NATS-Bench更容易运行。 此文件夹下每个文件的注释中列出了命令:https://github.com/D-X-Y/AutoDL-Projects/tree/main/exps/NATS-algos

— 您正在接收此消息,因为您编写了线程。 直接回复此邮件,在GitHub上查看,或取消订阅.

D-X-Y commented 3 years ago

You are welcome, I'm glad it helps!