Tathagat Dasgupta Tathagatd96

Senior Associate Consultant at Infosys

Tathagatd96 / gist:b22a13ddf55806815be6beb4610e5348

Created July 28, 2019 03:21

	trees_dump = bst.get_dump(fmap = "C:\\Users\\tatha\\.spyder-py3\\featmap.txt", with_stats = True)

	for trees in trees_dump:
	print(trees)


	xgb.plot_importance(bst, importance_type = 'gain', xlabel = 'Gain')

Tathagatd96 / gist:d1cc82ebec09be79a46f5eaf1f18b6a9

Last active July 28, 2019 03:13

	X_data=data.drop(["Class","Group"],axis=1)
	y_data=data["Class"]

	dtrain = xgb.DMatrix(X_data,y_data)

	params = {
	'objective':'binary:logistic',
	'max-depth':2,
	'silent':1,
	'eta':0.5

Tathagatd96 / gist:149c20d8df1e58d07598cb0bb0934c49

Created July 28, 2019 03:10

	X_data=data.drop(["Class","Group"],axis=1)
	y_data=data["Class"]

	dtrain = xgb.DMatrix(X_data,y_data)

	params = {
	'objective':'binary:logistic',
	'max-depth':2,
	'silent':1,
	'eta':0.5

Tathagatd96 / gist:017c0461d75c0a53664dd5301f528afb

Created July 28, 2019 03:03

	data=pd.read_csv("pima-indians-diabetes.csv")
	print(data.describe())
	print(data.keys())

	X_data=data.drop(["Class","Group"],axis=1)
	y_data=data["Class"]


	variable_params = {'max_depth':[2,4,6,10], 'n_estimators':[5, 10, 20, 25], 'learning_rate':np.linspace(1e-16, 1 , 3)}
	static_params = {'objective':'multi:softmax','num_class':4, 'silent':1}

Tathagatd96 / gist:1f79fedf4f074312581611b7f86418ce

Created July 28, 2019 00:19

	bst_grid = GridSearchCV (
	estimator = XGBClassifier(**static_params),
	param_grid = variable_params,
	scoring = "accuracy"
	)


	bst_grid.fit(X_data, y_data)

	print("Best Accuracy:{}".format(bst_grid.best_score_))

Tathagatd96 / gist:34f8e96067875d0190f6e1c158093ddd

Created July 28, 2019 00:18

	data=pd.read_csv("pima-indians-diabetes.csv")
	print(data.describe())
	print(data.keys())

	X_data=data.drop(["Class","Group"],axis=1)
	y_data=data["Class"]


	variable_params = {'max_depth':[2,4,6], 'n_estimators':[5, 10, 20, 25], 'learning_rate':np.linspace(1e-16, 1 , 3)}
	static_params = {'objective':'binary:logistic', 'silent':1}

Tathagatd96 / gist:2adf6b8a220bdc42fe2fd352d0c492bc

Created July 27, 2019 20:58

	import numpy as np
	import pandas as pd
	from xgboost.sklearn import XGBClassifier

	from sklearn.grid_search import RandomizedSearchCV
	from sklearn.cross_validation import StratifiedKFold
	import random
	import math

Tathagatd96 / gist:b424ee988ee4a8d0caf730701c499a8d

Created August 27, 2018 20:47

	samples=[] #generator examples

	with tf.Session() as sess:
	sess.run(init)
	for epoch in range(epochs):
	num_batches=mnist.train.num_examples//batch_size
	for i in range(num_batches):
	batch=mnist.train.next_batch(batch_size)
	batch_images=batch[0].reshape((batch_size,784))
	batch_images=batch_images*2-1

Tathagatd96 / gist:9c78dc23fa65da233da731cf7bc62972

Created August 27, 2018 20:42

	lr=0.001

	#Do this when multiple networks interact with each other
	tvars=tf.trainable_variables() #returns all variables created(the two variable scopes) and makes trainable true
	d_vars=[var for var in tvars if 'dis' in var.name]
	g_vars=[var for var in tvars if 'gen' in var.name]

	D_trainer=tf.train.AdamOptimizer(lr).minimize(D_loss,var_list=d_vars)
	G_trainer=tf.train.AdamOptimizer(lr).minimize(G_loss,var_list=g_vars)

Tathagatd96 / gist:d657bd0477d09097e065173706664868

Created August 27, 2018 20:32

	def loss_func(logits_in,labels_in):
	return tf.reduce_mean(tf.nn.sigmoid_cross_entropy_with_logits(logits=logits_in,labels=labels_in))

	D_real_loss=loss_func(D_logits_real,tf.ones_like(D_logits_real)*0.9) #Smoothing for generalization
	D_fake_loss=loss_func(D_logits_fake,tf.zeros_like(D_logits_real))
	D_loss=D_real_loss+D_fake_loss

	G_loss= loss_func(D_logits_fake,tf.ones_like(D_logits_fake))