Machine Learning and Deep Learning Models from Scratch.
This Library allows users to create the following models:

1. Feed-Forward Neural Networks
2. Convolution Neural Networks
3. Linear Regression
4. Logistic Regression

Without having to write any backpropagation code.

To install the Networks Library

pip install networks

No Params

No Params

No Params

No Params

batch_params={
  'mode':'train'/'test',
  'momentum':0.9,
  'eps':1e-8
  }

batch_params={
  'mode':'train'/'test',
  'momentum':0.9,
  'eps':1e-8
  }

pooling_params={
  'pooling_height':2,
  'pooling_width':2,
  'pooling_stride_height':2,
  'pooling_stride_width':2
}

num_kernels=64,
kernel_h=3,
kernel_w=3,
convolution_params={
  'stride':1
}

padding_h=2,
padding_w=2

No params

No params

No params

No params

affine_out = 64

No params

from networks.network import network
model = network(input_shape=(64,1,50,50),initialization="xavier2",
update_params={
  'alpha':1e-3,
  'method':'adam',
  'epoch':100,
  'reg':0.01,
  'reg_type':'L2',
  'offset':1e-7
})

model.add("padding",padding_h=3,padding_w=3)

model.add("convolution",num_kernels=64,kernel_h=3,kernel_w=3,
convolution_params:{
    'stride':1
  })

model.add("relu")

model.add("pooling",pooling_params={
  "pooling_height":2,
  "pooling_width":2,
  "pooling_stride_height":2,
  'pooling_stide_width':2
  })

model.add("batch_normalization",
batch_params={'mode':'train'/'test','momentum':0.9,'eps':1e-8})

model.add("spatial_batch",
batch_params={'mode':'train'/'test','momentum':0.9,'eps':1e-8})

model.add("flatten")

model.add("affine",affine_out=128)

model.add("softmax")

model.add("svm")

model.add("mse")

model.add("cross_entropy")

model.save("model.json")

model = network.load("model.json")

model.train(X,y)

accuracy,loss = model.test(validX,validY)

predictions = model.predict(X)