Last active
June 28, 2020 21:08
-
-
Save 2hands10fingers/ba2fb3b21788c5436241a99d110b0ffb to your computer and use it in GitHub Desktop.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
import numpy as np | |
""" | |
Special Notes: | |
- A model is your set weights and biases, and potentially an activation function | |
and potentially your type of network? Not sure. Need clarification | |
""" | |
# Inputs / variables from your dataset | |
np.random.seed(0) | |
class Layer_Dense: | |
# Not first layer. First layer is inputs. | |
# This could be as many layers as you want theoretically. | |
# acitvation_function(inputs * weight + bias) // activation_function(y = mx + b) | |
def __init__(self, n_inputs, n_neurons): | |
self.weights = self.rando_weights(n_inputs, n_neurons) | |
self.biases = self.bias_set(n_neurons) | |
def forward(self, inputs): | |
self.inputs = inputs | |
self.output = self.dot_prod(self.inputs, self.weights) + self.biases | |
# print(self.output) | |
if (self.activation): | |
self.activate() | |
return self | |
def bias_set(self, n_neurons): | |
return np.zeros((1, n_neurons)) | |
def rando_weights(self, n_inputs, n_neurons): | |
return 0.10 * np.random.randn(n_inputs, n_neurons) | |
def dot_prod(self, inputs, weights): | |
return np.dot(inputs, weights) | |
def activation_fn(self, acti_type): | |
activations = Activations() | |
self.activavtion_abrrev = acti_type | |
self.activation = activations.fn(acti_type) | |
return self | |
def activate(self): | |
if(self.output.any() and self.activation): | |
data = self.activation() | |
data.forward(self.output) | |
self.output = data.output | |
return self | |
def next(self, new=None): | |
neuro_length = len(self.output[0]) | |
next_handler = neuro_length if new == None else new | |
self.weights = self.rando_weights(neuro_length, next_handler) | |
self.bias = self.bias_set(neuro_length) | |
self.forward(self.output) | |
return self | |
def iterNext(self, layer_count): | |
for i in range(0,layer_count): | |
self.next() | |
return self | |
class Activations: | |
def __init__(self): | |
self.activations = { | |
"relu": Activation_relu | |
"sigmoid" Avtivation_sigmoid | |
} | |
def fn(self, act_type): | |
activation = act_type.lower() | |
self.active_activation = activation | |
return self.activations[activation] | |
class Activation_relu: | |
def forward(self, inputs): | |
self.output = np.maximum(0, inputs) | |
class Avtivation_sigmoid: | |
def forward(self, inputs): | |
# 1 / 1 + e ** -x | |
self.output = 1.0 / (1 + (2.718281828459045 * -x)) | |
class DataGenerator: | |
def spiral_data(self, points, classes): | |
X = np.zeros((points*classes, 2)) | |
y = np.zeros(points*classes, dtype='uint8') | |
for class_number in range(classes): | |
ix = range(points*class_number, points*(class_number+1)) | |
r = np.linspace(0.0, 1, points) # radius | |
t = np.linspace(class_number*4, (class_number+1)*4, points) + np.random.randn(points)*0.2 | |
X[ix] = np.c_[r*np.sin(t*2.5), r*np.cos(t*2.5)] | |
y[ix] = class_number | |
return X, y | |
# X = [[1,2,3,2.5],[2.0,5.0,-1.0,2.0],[-1.5,2.7,3.3,-0.8]] | |
gen = DataGenerator() | |
X, y = gen.spiral_data(100, 3) | |
layer_one = Layer_Dense(2,5) #(size, node/neuron count) | |
results = layer_one.activation_fn("relu").forward(X).iterNext(82) | |
print(results.output) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment