fusionlab 0.0.49

Useful packages for DL

FusionLab

FusionLab is an open-source frameworks built for Deep Learning research written in PyTorch and Tensorflow. The code is easy to read and modify especially for newbie. Feel free to send pull requests :D

• Installation
• How to use
• Encoders
• Segmentation
• Acknowledgements

Installation

With pip

pip install fusionlab

For Mac M1 chip users

Requirements:

• Apple Mac with M1 chips
• MacOS > 12.6 (Monterey)

Following steps

1. Clone this repo

git clone https://github.com/taipingeric/fusionlab.git
cd fusionlab

2. (remove anaconda first)
3. Install Miniconda
   1. Miniconda3 macOS Apple M1 64-bit pkg
   2. Miniconda3 macOS Apple M1 64-bit bash
4. Install the xcode-select command-line

xcode-select --install

5. Deactivate the base environment

conda deactivate 

6. Create conda environment using config

conda env create -f ./configs/tf-apple-m1-conda.yaml -n fusionlab

7. Replace requirements.txt with requirements-m1.txt
8. Install by pip

pip install -r requirements-m1.txt

ref: https://github.com/jeffheaton/t81_558_deep_learning/install/tensorflow-install-conda-mac-metal-jan-2023.ipynb

video

How to use

import fusionlab as fl

# PyTorch
encoder = fl.encoders.VGG16()
# Tensorflow
encoder = fl.encoders.TFVGG16()

Encoders

encoder list

Segmentation

import fusionlab as fl
# PyTorch UNet
unet = fl.segmentation.UNet(cin=3, num_cls=10, base_dim=64)

# Tensorflow UNet
import tensorflow as tf
# Binary Segmentation
unet = tf.keras.Sequential([
   fl.segmentation.TFUNet(num_cls=1, base_dim=64),
   tf.keras.layers.Activation(tf.nn.sigmoid),
])
unet.compile(loss=fl.losses.TFDiceLoss("binary"))

# Multiclass Segmentation
unet = tf.keras.Sequential([
   fl.segmentation.TFUNet(num_cls=10, base_dim=64),
   tf.keras.layers.Activation(tf.nn.softmax),
])
unet.compile(loss=fl.losses.TFDiceLoss("multiclass"))

Segmentation model list

• UNet, TFUNet
• ResUNet, TFResUNet
• UNet2plus, TFUNet2plus

Acknowledgements

• BloodAxe/pytorch-toolbelt