carefree-flow

Carefree?

carefree-learn aims to provide CAREFREE usages for both users and developers.

User Side

Computer Vision 🖼️

# MNIST classification task with LeNet

import cflow

import numpy as np
import oneflow.data as data


(x_train, y_train), (x_test, y_test) = data.load_mnist()
x_train, x_test = np.concatenate(x_train, axis=0), np.concatenate(x_test, axis=0)
y_train = np.concatenate(y_train, axis=0)[..., None]
y_test = np.concatenate(y_test, axis=0)[..., None]

data = cflow.cv.TensorData(x_train, y_train, x_test, y_test)
m = cflow.cv.CarefreePipeline(
    "clf",
    dict(
        in_channels=1,
        num_classes=10,
        img_size=28,
        latent_dim=128,
        encoder1d="lenet",
    ),
    fixed_epoch=5,
    loss_name="cross_entropy",
    metric_names=["acc", "auc"],
    tqdm_settings={"use_tqdm": True, "use_step_tqdm": True},
)
m.fit(data, cuda=0)

Developer Side

This is a WIP section :D

Installation

carefree-flow requires Python 3.6 or higher.

Pre-Installing OneFlow

carefree-flow requires oneflow>=0.4.0. Please refer to OneFlow for pre-installation.

pip installation

After installing OneFlow, installation of carefree-flow would be rather easy:

git clone https://github.com/carefree0910/carefree-flow
cd carefree-flow
pip install -e .

Citation

If you use carefree-flow in your research, we would greatly appreciate if you cite this library using this Bibtex:

@misc{carefree-flow,
  year={2021},
  author={Yujian He},
  title={carefree-flow, Deep Learning with OneFlow made easy},
  howpublished={\url{https://https://github.com/carefree0910/carefree-flow/}},
}

GitHub

GitHub - carefree0910/carefree-flow: Deep Learning ❤️ OneFlow
Deep Learning ❤️ OneFlow. Contribute to carefree0910/carefree-flow development by creating an account on GitHub.