forked from Yuejiang-li/info-diff
-
Notifications
You must be signed in to change notification settings - Fork 0
/
tsne.py
151 lines (115 loc) · 5.41 KB
/
tsne.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
import argparse
from sklearn.manifold import TSNE
import pickle
import numpy as np
import torch
import os
from loader import EmbLoader
from model import ModelWrapper
import matplotlib
matplotlib.use('Agg')
import seaborn as sns
import matplotlib.pyplot as plt
plt.style.use('seaborn')
parser = argparse.ArgumentParser()
parser.add_argument('--emb_file', type=str, default='./data/content_dict.pkl')
parser.add_argument('--model', type=str, default='Clash')
parser.add_argument('--idx_dict', type=str, default='./data/final/idx_dict.pkl')
parser.add_argument('--tsne_dump', type=str, default='./data/final/tsne_dump.pkl')
parser.add_argument('--key_id', type=int, default=113)
parser.add_argument('--cuda', type=bool, default=torch.cuda.is_available())
parser.add_argument('--data_dir', type=str, default='./data/final')
parser.add_argument('--followee_count_file', type=str, default='./data/followee_count.pkl')
parser.add_argument('--model_save_dir', type=str, default='./data/saved_model')
parser.add_argument('--save_epoch', type=int, default=5, help='Save model checkpoints every k epochs.')
parser.add_argument('--hidden_dim', type=int, default=200, help='RNN hidden state size.')
parser.add_argument('--num_layers', type=int, default=2, help='Num of RNN layers.')
parser.add_argument('--dropout', type=float, default=0.5, help='Input and RNN dropout rate.')
parser.add_argument('--penalty_coeff', type=float, default=0.5, help='Coefficient of Penalty term used in Clash model')
parser.add_argument('--num_epoch', type=int, default=30)
parser.add_argument('--log_step', type=int, default=20, help='Print log every k steps.')
parser.add_argument('--window_size', type=int, default=10)
parser.add_argument('--no_extra_linear', dest='use_extra_linear', action='store_false')
parser.add_argument('--use_extra_linear', dest='use_extra_linear', action='store_true')
parser.set_defaults(use_extra_linear=True)
parser.add_argument('--train_emb', dest='fix_emb', action='store_false')
parser.add_argument('--fix_emb', dest='fix_emb', action='store_true')
parser.set_defaults(fix_emb=False)
parser.add_argument('--patience', type=int, default=3)
parser.add_argument('--lr_decay', type=float, default=0.5)
parser.add_argument('--lr', type=float, default=1.0, help='Applies to SGD')
parser.add_argument('--max_grad_norm', type=float, default=5.0, help='Gradient clipping.')
parser.add_argument('--emb_dim', type=int, default=768)
parser.add_argument('--batch_size', type=int, default=256)
parser.add_argument('--seed', type=int, default=99)
parser.add_argument('--run_tsne', dest='tsne', action='store_true')
parser.set_defaults(tsne=False)
args = parser.parse_args()
opt = vars(args)
def cos_sim(a,b):
assert len(a) == len(b)
num = np.sum(np.multiply(a, b))
denom = np.linalg.norm(a) * np.linalg.norm(b) + 1e-10
cos = num / denom
return cos
with open(opt['emb_file'], 'rb') as fin:
emb = pickle.load(fin)
with open(opt['idx_dict'], 'rb') as fin:
weibo2embid = pickle.load(fin)
emb_matrix = EmbLoader(emb, weibo2embid, opt)
if opt['tsne']:
print('Original shape: ', np.shape(emb_matrix))
tsne = TSNE(n_components=2, random_state=0)
np.set_printoptions(suppress=True)
output = tsne.fit_transform(emb_matrix)
print('Current shape: ', np.shape(output))
with open(opt['tsne_dump'], 'wb') as fout:
pickle.dump(output, fout)
else:
with open(opt['tsne_dump'], 'rb') as fin:
output = pickle.load(fin)
key_emb = emb_matrix[opt['key_id'], :]
bert_sim = []
for i in range(len(emb_matrix)):
bert_sim.append(cos_sim(key_emb, emb_matrix[i, :]))
if not os.path.exists(os.path.join(opt['model_save_dir'], 'plot')):
os.mkdir(os.path.join(opt['model_save_dir'], 'plot'))
print('bert_sim max min: ', max(bert_sim), min(bert_sim))
fig = plt.figure()
cset = plt.scatter(output[:, 0], output[:, 1], s=8, c=bert_sim, cmap='PuBu')
plt.plot(output[opt['key_id'], 0], output[opt['key_id'], 1], 'r.')
plt.colorbar(cset)
plt.savefig(os.path.join(opt['model_save_dir'], 'plot', '%d_bert_sim' % opt['key_id']))
fig = plt.figure()
plt.hist(bert_sim, bins=100)
plt.savefig(os.path.join(opt['model_save_dir'], 'plot', '%d_bert_hist' % opt['key_id']))
model = ModelWrapper(opt, weibo2embid, eva=True)
model.load(os.path.join(opt['model_save_dir'], 'best_model.pt'))
clash_delta = []
for i in range(len(emb_matrix)):
emb1 = torch.tensor([i]).cuda()
emb2 = torch.tensor([opt['key_id']]).cuda()
clash_delta.append(model.model.get_delta(emb1, emb2).item())
ceil = max(max(clash_delta), min(clash_delta) * (-1)) * 1.05
print(max(clash_delta), min(clash_delta))
fig = plt.figure()
cset = plt.scatter(output[:, 0], output[:, 1], s=8, c=clash_delta, cmap='PuBu', vmin=-ceil, vmax=ceil)
plt.plot(output[opt['key_id'], 0], output[opt['key_id'], 1], 'r.')
plt.colorbar(cset)
plt.savefig(os.path.join(opt['model_save_dir'], 'plot', '%d_clash_delta' % opt['key_id']))
fig = plt.figure()
plt.hist(clash_delta, bins=100)
plt.savefig(os.path.join(opt['model_save_dir'], 'plot', '%d_clash_hist' % opt['key_id']))
c = np.polyfit(bert_sim, clash_delta, 1)
print(c)
k = c[0]
b = c[1]
label = 'y = {:.6f}x + {:.6f}'.format(k, b)
print(label)
fig = plt.figure()
assert len(clash_delta) == len(bert_sim)
sns.regplot(x=bert_sim, y=clash_delta, scatter_kws={'s':2}, label=label)
plt.legend(loc='upper right', fontsize='large')
plt.xlabel('Bert Cosine Similarity')
plt.ylabel('IMM Delta')
plt.savefig(os.path.join(opt['model_save_dir'], 'plot', '%d_scatter' % opt['key_id']))