smilelogging

Python logging package for easy reproducible experimenting in research.

Why you may need this package

This project is meant to provide an easy-to-use (as easy as possible) package to enable reproducible experimenting in research. Here is a struggling situation you may also encountered:

I am doing some project. I got a fatanstic idea some time (one week, one month, or even one year) ago. Now I am looking at the results of that experiment, but I just cannot reproduce them anymore. I cannot remember which script and what hyper-prarameters I used. Even worse, since then I’ve modified the code (a lot). I don’t know where I messed it up…

If you do not use this package, usually, what you can do may be:

  • First, use Github to manage your code. Always run experiments after git commit.
  • Second, before each experiment, set up a unique experiment folder (with a unique ID to label that experiment — we call it ExpID).
  • Third, when running an experiment, print your git commit ID (we call it CodeID) and arguments in the log.

Every result is uniquely binded with an ExpID, corresponding to a unique experiment folder. In that folder, CodeID and arguments are saved. So ideally, as long as we know the ExpID, we should be able to rerun the experiment under the same condition.

These steps are pretty simple, but if you implement them over and over again in each project, it can still be quite annoying. This package is meant to save you with basically 3~4 lines of code change.

Usage

Step 0: Install the package (>= python3.4)

# --upgrade to make sure you install the latest version
pip install smilelogging --upgrade

Step 1: Modify your code

Here we use the official PyTorch ImageNet example to give an example.

# 1. add this at the head of code
from smilelogging import Logger 

# 2. replace argument parser
parser = argparse.ArgumentParser(description='PyTorch ImageNet Training')  
==> 
from smilelogging import argparser as parser

# 3. add logger and change print if necessary
args = parser.parse_args()
==> 
args = parser.parse_args()
logger = Logger(args)
global print; print = logger.log_printer.logprint # change print function so that logs can be printed to a txt file

TIPS: overwriting the default python print func may not be a good practice, a better way may be logprint = logger.log_printer.logprint, and use it like logprint('Test accuracy: %.4f' % test_acc). This will print the log to a txt file at path log/log.txt.

Step 2: Run experiments

The original ImageNet training snippet is:

CUDA_VISIBLE_DEVICES=0 python main.py -a resnet18 [imagenet-folder with train and val folders]

Now, try this:

CUDA_VISIBLE_DEVICES=0 python main.py -a resnet18 [imagenet-folder with train and val folders] --project_name Scratch__resnet18__imagenet --screen_print

This snippet will set up an experiment folder under path Experiments/Scratch__resnet18__imagenet_XXX. That XXX thing is an ExpID automatically assigned by the time running this snippet. Below is an example on my PC:

Experiments/
└── Scratch__resnet18__imagenet_SERVER138-20211021-145936
    ├── gen_img
    ├── log
    │   ├── git_status.txt
    │   ├── gpu_info.txt
    │   ├── log.txt
    │   ├── params.yaml
    │   └── plot
    └── weights

✨Congrats❗You’re all set❗

As seen, there will be 3 folders automatically created: gen_img, weights, log. Log text will be saved in log/log.txt, arguments saved in log/params.yaml and in the head of log/log.txt. Below is an example of the first few lines of log/log.txt:

cd /home/wanghuan/Projects/TestProject
CUDA_VISIBLE_DEVICES=1 python main.py -a resnet18 /home/wanghuan/Dataset/ILSVRC/Data/CLS-LOC/ --project Scracth_resnet18_imagenet --screen_print

('arch': resnet18) ('batch_size': 256) ('cache_ignore': ) ('CodeID': f30e6078) ('data': /home/wanghuan/Dataset/ILSVRC/Data/CLS-LOC/) ('debug': False) ('dist_backend': nccl) ('dist_url': tcp://224.66.41.62:23456) ('epochs': 90) ('evaluate': False) ('gpu': None) ('lr': 0.1) ('momentum': 0.9) ('multiprocessing_distributed': False) ('note': ) ('pretrained': False) ('print_freq': 10) ('project_name': Scracth_resnet18_imagenet) ('rank': -1) ('resume': ) ('screen_print': True) ('seed': None) ('start_epoch': 0) ('weight_decay': 0.0001) ('workers': 4) ('world_size': -1)

[180853 22509 2021/10/21-18:08:54] ==> Caching various config files to 'Experiments/Scracth_resnet18_imagenet_SERVER138-20211021-180853/.caches'

Note, it tells us

  • (1) where is the code
  • (2) what snippet is used when running this experiment
  • (3) what arguments are used
  • (4) what is the CodeID — useful when rolling back to prior code versions (git reset --hard <CodeID>)
  • (5) where the code files (*.py, *.json, *.yaml etc) are backuped — note the log line “==> Caching various config files to …”. Ideally, CodeID is already enough to get previous code. Caching code files is a double insurance
  • (6) At the begining of each log line, the prefix “[180853 22509 2021/10/21-18:08:54]” is automatically added if the logprint func is used for print, where 180853 is short for the full ExpID SERVER138-20211021-180853, 22509 is the program pid (useful if you want to kill the job, e.g., kill -9 22509)

More explanantions about the folder setting

The weights folder is supposed to store the checkpoints during training; and gen_img is supposed to store the generated images during training (like in a generative model project). To use them in the code:

weights_path = logger.weights_path
gen_img_path = logger.gen_img_path

More explanantions about the arguments and more tips

  • --screen_print means the logs will also be print to the console (namely, your screen). If it is not used, the log will only be saved to log/log.txt, not printed to screen.
  • If you are debugging code, you may not want to create an experiment folder under Experiments. Then use --debug, for example:

CUDA_VISIBLE_DEVICES=0 python main.py -a resnet18 [imagenet-folder with train and val folders] --debug

This will save all the logs in Debug_Dir, instead of Experiments (Experiments is expected to store the formal experiment results).

TODO

  • Add training and testing metric (like accuracy, PSNR) plots.

Collaboration / Suggestions

Currently, this is still a baby project. Any collaboration or suggestions are welcome to Huan Wang (Email: [email protected]).

GitHub

View Github