def mbgd():
X, X_val, y, y_val = data.demo()
theta = np.array([-1,-1])
learning_rate = 0.01
learning_rate_sched = 't'
time_decay = 0.1
step_decay=0.5
step_epochs=2
exp_decay=0.1
precision = 0.001
maxiter = 10000
directory = "./test/figures/BGD/"
i_s=5
batch_size = 5
gd = MBGD()
gd.fit(X=X, y=y, theta=theta,X_val=X_val, y_val=y_val, learning_rate=learning_rate,
learning_rate_sched=learning_rate_sched, time_decay=time_decay, step_decay=step_decay,
step_epochs=step_epochs, exp_decay=exp_decay,
maxiter=maxiter, precision=precision, i_s=i_s)
return(gd)
# %% import numpy as np import pandas as pd import sys srcdir = "c:\\Users\\John\\Documents\\Data Science\\Libraries\\GradientDescent\\src" sys.path.append(srcdir) from GradientDescent import BGD, SGD from GradientVisual import GradientVisual import data #%% # --------------------------------------------------------------------------- # # SEARCH FUNCTION # # --------------------------------------------------------------------------- # X, X_val, y, y_val = data.demo() alg = 'Stochastic Gradient Descent' theta = np.array([-1, -1]) learning_rate = 0.01 learning_rate_sched = 'c' stop_metric = 'j' time_decay = 0.5 step_decay = 0.5 step_epochs = 2 exp_decay = 0.1 precision = 0.01 maxiter = 5000 i_s = 5 directory = "./test/figures/SGD/" gd = SGD() gd.fit(X=X,
import os import sys from IPython.display import HTML from matplotlib import animation, rc from matplotlib import rcParams import numpy as np import pandas as pd src = "c:\\Users\\John\\Documents\\Data Science\\Libraries\\GradientDescent\\src" sys.path.append(src) from GradientLab import BGDLab import data # Data X, X_val, y, y_val = data.demo(n=500) # Parameters theta = np.array([-1, -1]) learning_rate = [1.6] stop_metric = ['j'] precision = [0.001] maxiter = 5000 learning_rate_sched = ['t'] time_decay = np.arange(0.01, 1, 0.01) step_decay = [0.1, 0.01, 0.001] step_epochs = [2, 5, 10] exp_decay = [0.1, 0.01, 0.001] i_s = [10] directory = "./test/figures/BGD/Lab/"