Classifier-trains

PyPI version Python version Downloads license CI

Install as a package

pip3 install classifier-trains

Examples of Pipeline Configuration

For example of pipeline configuration, please see pipeline_config_only_train.yml, pipeline_config_only_eval.yml, full_pipeline_config.yml.

# Run training or evaluation
python3 -m classifier_trains run --config <path to yml config> --output_dir <output_dir>

# Run training or evaluation with profiling, which will generate a profile report
python3 -m classifier_trains profile --config <path to yml config> --output_dir <output_dir>

# Compute mean and std of dataset
python3 -m classifier_trains compute-mean-and-std --dir-path <path to dataset>

# Get output mapping of dataset
python3 -m classifier_trains get-output-mapping --dir-path <path to dataset>

Expected Folder Structure for Dataset

Dataset Directory
├── train
│   ├── <class1>
│   │   ├── <image1>
│   │   └── ...
│   └── <class2>
│       ├── <image1>
│       └── ...
├── val
│   ├── <class1>
│   │   ├── <image1>
│   │   └── ...
│   └── <class2>
│       ├── <image1>
│       └── ...
└── eval
    ├── <class1>
    │   ├── <image1>
    │   └── ...
    └── <class2>
        ├── <image1>
        └── ...

Pipeline Parameter Table

Pipeline Parameters

Parameter Description Type Default Choices
enable_training Enable training bool False True, False
enable_evaluation Enable evaluation bool False True, False

Model Parameters

Parameter Description Type Default Choices
model Model architecture to use str / resnet18, resnet34, resnet50, resnet152, vgg11, vgg11_bn, vgg13, vgg13_bn, vgg16, vgg16_bn, vgg19, vgg19_bn, squeezenet1_0, squeezenet1_1, densenet121, densenet161, densenet169, densenet201, inception_v3
num_classes Number of classes int / Any positive integer
weights Pretrained weights str DEFAULT Check PyTorch Documentation
checkpoint_path Path to checkpoint Optional[str] None Self defined path

Dataloader Parameters

Parameter Description Type Default Choices
batch_size Batch size int / Any positive integer
num_workers Number of workers int 0 Any non negative integer

Resize Parameters

Parameter Description Type Default Choices
width Width of resized image int / Any positive integer
height Height of resized image int / Any positive integer
interpolation Interpolation method str bicubic nearest, nearest_exact, bilinear, bicubic
padding Padding Optional[str] None top_left, top_right, bottom_left, bottom_right, center
maintain_aspect_ratio Maintain aspect ratio bool False True, False

Spatial Transform Parameters

Parameter Description Type Default Choices
hflip_prob Horizontal flip probability float / Any float between 0 and 1
vflip_prob Vertical flip probability float / Any float between 0 and 1
max_rotate_in_degree Maximum rotation in degree float 0.0 Any non negative float
allow_center_crop Allow center crop bool False True, False
allow_random_crop Allow random crop bool False True, False

Color Transform Parameters

Parameter Description Type Default Choices
allow_gray_scale Allow gray scale bool False True, False
allow_random_color Allow random color bool False True, False

Preprocessing Parameters

Parameter Description Type Default Choices
mean Mean of dataset List[float] / Any list of floats
std Standard deviation of dataset List[float] / Any list of floats

Optimizer Parameters

Parameter Description Type Default Choices
name Optimizer to use str / sgd, adam, rmsprop, adamw
lr Learning rate float / Any positive float
weight_decay Weight decay float 0.0 Any non negative float
momentum Momentum Optional[float] None Any non negative float
alpha Alpha Optional[float] None Any non negative float
betas Betas Optional[Tuple[float, float]] None Any tuple of non negative floats

Scheduler Parameters

Parameter Description Type Default Choices
name Scheduler to use str / step, consine
lr_min Minimum learning rate Optional[float] None Any non negative float
step_size Step size Optional[int] None Any positive integer
gamma Gamma Optional[float] None Any non negative float

Training Parameters

Parameter Description Type Default Choices
name Name of the experiment str / Any string
num_epochs Number of epochs int / Any positive integer
trainset_dir Path to training dataset str / Any string
valset_dir Path to validation dataset str / Any string
testset_dir Path to testing dataset Optional[str] None Any string
device Device to use str cuda cpu, cuda, etc.
max_num_hrs Maximum number of hours to run Optional[float] None Any non negative float
criterion What criterion to use for best model str loss loss, accuracy
validate_every Validate every n epochs int 1 Any positive integer
save_every Save model every n epochs int 3 Any positive integer
patience Patience for early stopping int 5 Any positive integer
random_seed Random seed int 42 Any positive integer
precision Precision int 64 16, 32, 64
export_last_as_onnx Export last model as ONNX bool False True, False
export_best_as_onnx Export best model as ONNX bool False True, False

Evaluation Parameters

Parameter Description Type Default Choices
name Name of the experiment str / Any string
evalset_dir Path to evaluation dataset str / Any string
device Device to use str cuda cpu, cuda, etc.
precision Precision int 64 16, 32, 64
random_seed Random seed int 42 Any positive integer

Project links

Data verified by PyPI on Jan 10, 2026
Data provided by the project maintainers, verified at the time the release was uploaded to PyPI.

Key dates

PyPI data
Data sourced directly from PyPI's database.
  • Released:
Latest release

1 maintainer

PyPI data
Data sourced directly from PyPI's database.
Avatar for wyhwong from gravatar.com wyhwong

Credits

Author: wyhwong

License

MIT License (MIT)

Requires

Python <4.0, >=3.10

Tags

pytorch deep learning image classification

Classifiers

Report project as malware

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

classifier_trains-1.2.4.tar.gz (19.5 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

classifier_trains-1.2.4-py3-none-any.whl (25.2 kB view details)

Uploaded Python 3

File details

Details for the file classifier_trains-1.2.4.tar.gz.

File metadata

  • Download URL: classifier_trains-1.2.4.tar.gz
  • Upload date:
  • Size: 19.5 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: poetry/1.8.4 CPython/3.11.14 Linux/6.8.0-1044-azure

File hashes

Hashes for classifier_trains-1.2.4.tar.gz
Algorithm Hash digest
SHA256 4c53bf1b528b339b660b675e459a99d79d36de49574134a892c5f97106667a71
MD5 8345ec20b4e5cfacf1df4bf3b0648a13
BLAKE2b-256 549d646625ed1e272aa8bb94baf4944ca243cb879b493888c84579713f8337fc

See more details on using hashes here.

File details

Details for the file classifier_trains-1.2.4-py3-none-any.whl.

File metadata

  • Download URL: classifier_trains-1.2.4-py3-none-any.whl
  • Upload date:
  • Size: 25.2 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: poetry/1.8.4 CPython/3.11.14 Linux/6.8.0-1044-azure

File hashes

Hashes for classifier_trains-1.2.4-py3-none-any.whl
Algorithm Hash digest
SHA256 4d4cfd96b4f0a43a835b6ff6119f3ffd9719948db3c37fb2b104850fb842dac8
MD5 b36ca6e9f51bd8ddcc625e6c56a64497
BLAKE2b-256 598e16025b39102d99cb0184c88755af118eed707f45eaa5717057892197d8f6

See more details on using hashes here.

Release history Release notifications | RSS feed

This release

1.2.4 This release

2 files

1.2.3

2 files

1.2.2

2 files

1.2.1

2 files

1.2.0

2 files

1.1.9

2 files

1.1.8

2 files

1.1.7

2 files

1.1.6

2 files

1.1.5

2 files

1.1.4

2 files

1.1.3

2 files

1.1.2

2 files

1.1.1

2 files

1.1.0

2 files

1.0.0

2 files

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page