1

使用aim工具来记录机器学习训练

 2 years ago
source link: https://juzq.github.io/posts/aim_tool_record_train_runs/
Go to the source link to view the article. You can view the picture content, updated content and better typesetting reading experience. If the link is broken, please click the button below to view the snapshot at that time.
neoserver,ios ssh client
Posts 使用aim工具来记录机器学习训练

使用aim工具来记录机器学习训练

Jan 18, 20212021-01-19T00:00:00+08:00 by JuZi

aim能够非常简单方便地记录、搜索、对比机器学习训练的超参数。相比于tensorboard，它支持将多个指标合并到一张图上，并且支持类似sql的方式来搜索和对比超参数，可操控性会更加精细。

  1. 安装aim：pip install aim
  2. 安装docker：sudo apt install docker.io

初始化aim仓库

命令：aim init

通过该命令，aim会在当前目录创建.aim文件夹，用于记录后续的训练数据，因此必须在数据记录前初始化aim仓库。

数据记录示例

import aim
lrs = [1e-2,1e-3,1e-4]
bss = [64, 128, 256, 512]
layer_nums = [1, 2, 3]

x = 1
for lr in lrs:
    for bs in bss:
        for ln in layer_nums:
            hyperparam_dict={"lr":lr,"batch_size":bs,"layer_num":ln}
            sess = aim.Session(experiment='test')
            sess.set_params(hyperparam_dict, name='hparams')
            for epoch_number in range(1,10):
                sess.track(x/epoch_number+1, name='frames', step=epoch_number)
                sess.track(x/epoch_number, name='loss', step=epoch_number)
            x+=1

启动UI界面

命令：aim up -h 192.168.102.65 -p 9901

若不指定-h和-p参数，aim会默认使用127.0.0.1:43800作为监听地址。

This post is licensed under CC BY 4.0 by the author.
Share

report

Recommend

About Joyk


Aggregate valuable and interesting links.
Joyk means Joy of geeK