This package is a light-weight core library that provides the most common and essential functionalities shared in various deep learning tasks:
Trainer: does tedious training logic for you.LRWarmupScheduler: wraps all standard PyTorch LR scheduler to support warmup.ConfigArgumentParser: provides an argument parser that supports loading a YAML configuration file.- ......
You can find a brief Chinese introduction at zhihu.
[2023/10/29]: Now CPU supports iteration-based training, checkpointing and evaluation!
From PyPI.
pip install core-pytorch-utils
Or from source.
git clone https://github.com/serend1p1ty/core-pytorch-utils.git
cd core-pytorch-utils
pip install -r requirements.txt
pip install -v -e .
In examples/ folder, we show how to use our Trainer to train a CNN on MINIST.
It is strongly recommended that you run this code before using the CPU library.
Learn more from our documentaion.
Pull request is welcomed! Before submitting a PR, DO NOT forget to run ./dev/linter.sh that provides syntax checking and code style optimation.
CPU is released under the MIT License.
We refered mmcv, detectron2 and pytorch-image-models when develping CPU.