-
Notifications
You must be signed in to change notification settings - Fork 6
/
VAE_rec_main.py
180 lines (140 loc) · 5.9 KB
/
VAE_rec_main.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
# -*- coding: utf-8 -*-
"""
Created on Tue Mar 22 10:43:29 2016
@author: Rob Romijnders
TODO
- Cross validate over different learning-rates
"""
import sys
import socket
if 'rob-laptop' in socket.gethostname():
sys.path.append('/home/rob/Dropbox/ml_projects/basket_local/')
sys.path.append('/home/rob/Dropbox/ml_projects/basket_local/SportVU-seq')
#The folder where your dataset is. Note that is must end with a '/'
direc = '/home/rob/Dropbox/ml_projects/basket_local/SportVU-seq/'
elif 'rob-com' in socket.gethostname():
sys.path.append('/home/rob/Documents/nn_sportvu')
direc = '/home/rob/Documents/nn_sportvu/SportVU-seq/'
#Rajiv: you can add your computer name here
import numpy as np
import tensorflow as tf
import matplotlib.pyplot as plt
from tensorflow.python.framework import ops
from tensorflow.python.ops import clip_ops
from basket_util import *
from VAE_util import *
import sklearn as sk
from sklearn.metrics import roc_auc_score,roc_curve
from data_loader_class import *
from mpl_toolkits.mplot3d import axes3d
import matplotlib.mlab as mlab
from VAE_rec_model_reverse import *
"""Hyperparameters"""
config = {}
config['num_layers'] = 2
config['hidden_size'] = 60
config['max_grad_norm'] = 1
config['batch_size'] = batch_size = 64
config['sl'] = sl = 18 #sequence length
config['mixtures'] = 1
config['learning_rate'] = .005
config['num_l'] = num_l = 2
ratio = 0.8 #Ratio for train-val split
plot_every = 100 #How often do you want terminal output for the performances
max_iterations = 50000
"""Load the data"""
#The name of the dataset. Note that it must end with '.csv'
csv_file = 'seq_all_9feet.csv'
#Load an instance
center = np.array([5.25, 25.0, 10.0])
dl = DataLoad(direc,csv_file, center)
#Munge the data. Arguments see the class
db = 4 #distance to basket
dl.munge_data(11,sl,db)
#Center the data
dl.center_data(center)
dl.entropy_offset()
dl.split_train_test(ratio = 0.8)
data_dict = dl.data
dl.plot_traj_2d(20,'at %.0f feet from basket'%db)
X_train = np.transpose(data_dict['X_train'],[0,2,1])
#y_train = data_dict['y_train']
X_val = np.transpose(data_dict['X_val'],[0,2,1])
y_val = data_dict['y_val']
N,crd,_ = X_train.shape
Nval = X_val.shape[0]
config['crd'] = crd
#Proclaim the epochs
epochs = np.floor(batch_size*max_iterations / N)
print('Train with approximately %d epochs' %(epochs))
model = Model(config)
# For now, we collect performances in a Numpy array.
# In future releases, I hope TensorBoard allows for more
# flexibility in plotting
perf_collect = np.zeros((7,int(np.floor(max_iterations /plot_every))))
sess = tf.Session()
#with tf.Session() as sess:
if True:
writer = tf.train.SummaryWriter("/home/rob/Dropbox/ml_projects/basket_local/nn_sportvu/log_tb", sess.graph)
sess.run(tf.initialize_all_variables())
step = 0 # Step is a counter for filling the numpy array perf_collect
for i in range(max_iterations):
batch_ind = np.random.choice(N,batch_size,replace=False)
# debug = sess.run(model.sl_t,feed_dict={model.x:X_train[batch_ind], model.y_: y_train[batch_ind], model.keep_prob: dropout})
# print(np.max(debug[0]))
# print(np.max(debug[1]))
if i%plot_every == 0:
#Check training performance
fetch = [model.cost_seq, model.cost_kld, model.cost_xstart]
result = sess.run(fetch,feed_dict = { model.x: X_train[batch_ind]})
perf_collect[0,step] = cost_train_seq = result[0]
perf_collect[1,step] = cost_train_kld = result[1]
perf_collect[4,step] = cost_train_xstart = result[2]
#Check validation performance
batch_ind_val = np.random.choice(Nval,batch_size,replace=False)
fetch = [model.cost_seq, model.cost_kld, model.cost_xstart] #, model.merged
result = sess.run(fetch, feed_dict={ model.x: X_val[batch_ind_val]})
perf_collect[2,step] = cost_val_seq = result[0]
perf_collect[3,step] = cost_val_kld = result[1]
perf_collect[5,step] = cost_val_xstart = result[2]
# #Write information to TensorBoard
# summary_str = result[3]
# writer.add_summary(summary_str, i)
# writer.flush() #Don't forget this command! It makes sure Python writes the summaries to the log-file
print("At %6s / %6s train (%6.3f,%6.3f,%6.3f) val (%6.3f,%6.3f,%6.3f)" % (i,max_iterations,cost_train_seq,cost_train_kld,cost_train_xstart,cost_val_seq,cost_val_kld,cost_val_xstart ))
step +=1
sess.run(model.train_step,feed_dict={model.x:X_train[batch_ind]})
#In the next line we also fetch the softmax outputs
batch_ind_val = np.random.choice(Nval,batch_size,replace=False)
result = sess.run([model.numel], feed_dict={ model.x: X_val[batch_ind_val]})
print('The network has %s trainable parameters'%(result[0]))
# debug = sess.run(model.b_xend)
z_feed = np.random.randn(batch_size,num_l)
result = sess.run(model.x_col, feed_dict={ model.z: z_feed})
X_vae = np.transpose(result,[1,2,0])
labels_dummy = np.random.randint(0,1,size=(batch_size,1))
plot_basket(X_vae,labels_dummy)
"""Visualize the 2D latent space"""
label_type = 'class' #Color scatter plot according to hit/miss
label_type = 'x' #Color the scatter plot according to x coordinate
label_type = 'y' #Color scatter plot according to y coordinate
if num_l == 2:
##Extract the latent space coordinates of the validation set
start = 0
label = [] #The label to save to visualize the latent space
z_run = []
while start + batch_size < Nval:
run_ind = range(start,start+batch_size)
z_mu_fetch = sess.run(model.z_mu, feed_dict = {model.x:X_val[run_ind]})
z_run.append(z_mu_fetch)
if label_type == 'y':
label.append(X_val[run_ind,1,0]) #The y coordinate of x_start
if label_type == 'x':
label.append(X_val[run_ind,0,0]) #The y coordinate of x_start
if label_type == 'class':
label.append(y_val[run_ind])
start += batch_size
z_run = np.concatenate(z_run,axis=0)
label = np.concatenate(label,axis=0)
plt.figure()
plt.scatter(z_run[:,0],z_run[:,1],c = label,linewidths=0.0)