HelixNet documentation

Welcome to HelixNet documentation.

A small, extensible, lightweight and powerful building blocks that can build amazing stuff without much hassle, complexity and can run on a potato hardware.

1. Lightweight and Fast

2. No need to compile 2000+ C++ files.

3. No need to clone 1GB+ Repos.

Just plug and play.

Contents:

• Installation
  • Using Pip
  • Installation for Development
• Optimizers
  • Optimizer
  • SGD
  • Adam
  • RMSProp
  • NesterovSGD
• Learn Rates
  • LearnRate
  • ExpDecay
  • LinearDecay
• Regularizers
  • Regularizer
  • L1
  • L2
• Layers
  • Layer
  • Dense
  • Conv2D
  • MaxPooling2D
  • Flatten
  • Reshape
  • Embedding
  • InputShape
  • LSTMLayer
  • LSTMCell
  • BatchNorm
  • Dropout
  • DenseTranspose
  • ConvTranspose2D
• Losses
  • BinaryCrossEntropy()
  • CatCrossEntropy()
  • CosineSimilarityLoss()
  • FocalLoss()
  • HingeLoss()
  • HuberLoss()
  • Kullback_Leibler()
  • LogCoshLoss()
  • MeanAbsError()
  • MeanSquaredError()
  • softmax_crossentropy()
• Metrics
  • accuracy()
  • f1_score()
  • precision()
  • r_squared()
  • recall()
  • rmse()
  • top_k_accuracy()
• Models
  • Sequential
• Preprocessing
  • load_image()
• Input and Output File Operations
  • save_model()
  • load_model()
• DIY: Extending HelixNet
  • Optimizers
• Changelog
  • New in 0.7.0
  • New in 0.6.1
  • New in 0.6.0
  • New in 0.5.1
  • New in 0.5.0
  • New in 0.4.0
  • New in 0.3.0