wk-classify

A package of tools for building deep-learning classification programs. Easy to use, light and powerful.

Install

pip3 install wk-classify

Usage

Quick start

from wcf import train,  TrainValConfigBase
class Config(TrainValConfigBase):
    TRAIN_DIR = 'path for train set'
    VAL_DIR = 'path for val set'
cfg=Config()
train(cfg)

A real example

from wcf import train, TrainValConfigBase, val,t,EasyTransform,models_names
class Config(TrainValConfigBase):
    MODEL_TYPE = models_names.shufflenet_v2_x0_5
    TAG = '[%s]'%(MODEL_TYPE)
    GEN_CLASSES_FILE = True
    USE_tqdm_TRAIN = True
    INPUT_SIZE = (252,196) #(w,h)
    BATCH_SIZE = 64
    MAX_EPOCHS = 200
    BALANCE_CLASSES = True
    VAL_INTERVAL = 1
    WEIGHTS_SAVE_INTERVAL = 1
    WEIGHTS_INIT = 'weights/training/model_best.pkl'
    TRAIN_DIR = '/home/ars/sda5/data/projects/烟盒/data/现场采集好坏烟照片/相机1-train'
    VAL_DIR = '/home/ars/sda5/data/projects/烟盒/data/现场采集好坏烟照片/相机1-val'
    val_transform = EasyTransform([
        t.Resize(INPUT_SIZE[::-1]),
        t.SaveToDir('data/test'),
        t.ToTensor(),
    ])
    train_transform = EasyTransform([
        t.ColorJitter(brightness=0.2, contrast=0, saturation=0, hue=0),
        # t.RandomHorizontalFlip(),
        # t.RandomVerticalFlip(),
        # t.RandomRotate(360),
        t.RandomTranslate(30),
        t.RandomBlur(p=0.3, radius=1),
        t.RandomSPNoise(p=0.3),
        *val_transform,
    ])
    # def get_model(self, num_classes=None):
        # model=YourModel(...)
        # return model


if __name__ == '__main__':
    cfg = Config()
    train(cfg)
    # res=val(cfg)
    # print(res)

all options

see the TrainValConfigBase class for all options

how to predict?

see demo_predict.py

see demo_train.py and demo_predict.py

License
- OSI Approved :: MIT License
Operating System
- OS Independent
Programming Language
- Python :: 3

Release files for wk-classify 0.0.3.5

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for wk-classify 0.0.3.5
File	Size	Uploaded
wk-classify-0.0.3.5.tar.gz	24.3 kB	Sep 24, 2020	Details

Built distribution (wheel)

Table of built distributions (wheels) for wk-classify 0.0.3.5
File	Interpreter	ABI	Platform	Reset
wk_classify-0.0.3.5-py3-none-any.whl 31.3 kB Sep 24, 2020	Python 3	none	any	Details

Total release size: 55.6 kB

Release files / wk-classify-0.0.3.5.tar.gz

Download URL	wk-classify-0.0.3.5.tar.gz
Size	24.3 kB
Tags	Source
SHA-256 checksum How to use checksums	`5d0e92fadda9b9f71db35bdd1aceaddd4348c2ad2f69d29b61bd7521ae66270e`
BLAKE2b-256 checksum How to use checksums	`57b108565928bb24c6991b6e154271b6544aa8190669f44817a362993adb0d8d`
Upload date	Sep 24, 2020
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`twine/3.1.1 pkginfo/1.5.0.1 requests/2.24.0 setuptools/47.3.1 requests-toolbelt/0.9.1 tqdm/4.47.0 CPython/3.6.9`

Release files / wk_classify-0.0.3.5-py3-none-any.whl

Download URL	wk_classify-0.0.3.5-py3-none-any.whl
Size	31.3 kB
Tags	Python 3
SHA-256 checksum How to use checksums	`4294fcd0b53e6dd45fc1fdb51d0b3fb438dc4141d1614accd6357eabb0f0dfb1`
BLAKE2b-256 checksum How to use checksums	`8e1fcd1792e7d505d3d6607383ac6082e6981776957c2bba76512b7defc6df7a`
Upload date	Sep 24, 2020
Uploaded using Trusted Publishing? What is trusted publishing?	No
Uploaded via	`twine/3.1.1 pkginfo/1.5.0.1 requests/2.24.0 setuptools/47.3.1 requests-toolbelt/0.9.1 tqdm/4.47.0 CPython/3.6.9`