-
Notifications
You must be signed in to change notification settings - Fork 0
/
interact_flask.py
220 lines (178 loc) · 6.08 KB
/
interact_flask.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
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
# -*- coding: utf-8 -*-
from transformers.modeling_gpt2 import GPT2LMHeadModel
from transformers import GPT2Tokenizer
from pplm_interactive import generate_text_pplm
import torch
import numpy as np
import json
from typing import List, Optional, Tuple, Union
from pplm_classification_head import ClassificationHead
from transformers.file_utils import cached_path
import os
import flask
from flask import request
import argparse
#This is done to fix an issue on mac os with xgboost and matplotlib. Please comment this line if using in any other OS
os.environ['KMP_DUPLICATE_LIB_OK']='True'
app = flask.Flask(__name__)
#app.config["DEBUG"] = True
model_size = "medium"
pretrained_model="DialoGPT-"+model_size
discrim_meta_file="generic_classifier_head_meta.json"
discrim_weights="generic_classifier_head.pt"
PPLM_DISCRIM = 2
loss_type = PPLM_DISCRIM
def get_classifier(
discrim_meta: Optional[dict],
class_label: Union[str, int],
device: str
) -> Tuple[Optional[ClassificationHead], Optional[int]]:
if discrim_meta is None:
return None, None
params = discrim_meta
classifier = ClassificationHead(
class_size=params['class_size'],
embed_size=params['embed_size']
).to(device)
if "url" in params:
resolved_archive_file = cached_path(params["url"])
elif "path" in params:
resolved_archive_file = params["path"]
else:
raise ValueError("Either url or path have to be specified "
"in the discriminator model parameters")
classifier.load_state_dict(
torch.load(resolved_archive_file, map_location=device))
classifier.eval()
if isinstance(class_label, str):
if class_label in params["class_vocab"]:
label_id = params["class_vocab"][class_label]
else:
label_id = params["default_class"]
elif isinstance(class_label, int):
if class_label in set(params["class_vocab"].values()):
label_id = class_label
else:
label_id = params["default_class"]
else:
label_id = params["default_class"]
return classifier, label_id
# set Random seed
pert_gen_tok_texts = []
discrim_losses = []
losses_in_time = []
@app.route('/', methods=['GET'])
def home():
return '''<h1>Dialogue API</h1>
<p>A dialogue API based on a transformer model.</p>'''
output_so_far=None
@app.route('/api/getresponse', methods=['GET'])
def getresponse():
global output_so_far
cond_text= request.args.get('query')
context = tokenizer.encode(tokenizer.bos_token + cond_text +tokenizer.bos_token,add_special_tokens=False)
if context:
context_t = torch.tensor(context, device=device, dtype=torch.long)
while len(context_t.shape) < 2:
context_t = context_t.unsqueeze(0)
output_so_far = (
context_t if output_so_far is None
else torch.cat((output_so_far, context_t), dim=1)
)
min_discrim_loss=np.infty
selected_pert_gen_tok_text=None
for i in range(num_samples):
pert_gen_tok_text, discrim_loss, loss_in_time = generate_text_pplm(
model=model,
tokenizer=tokenizer,
output_so_far=output_so_far,
device=device,
perturb=True,
classifier=classifier,
class_label=class_id,
loss_type=loss_type,
length=200,
stepsize=0.05,
temperature=1.0,
top_k=10,
sample=True,
num_iterations=2,
grad_length=10000,
horizon_length=1,
window_length=0,
decay=False,
gamma=1,
gm_scale=0.8,
kl_scale=0.01
)
dl=discrim_loss.data.cpu().numpy().tolist()
#print('resp: {} , loss: {}'.format(tokenizer.decode(pert_gen_tok_text.tolist()[0]),dl))
if(dl<min_discrim_loss):
min_discrim_loss=dl
selected_pert_gen_tok_text=pert_gen_tok_text
output_so_far = (
selected_pert_gen_tok_text if output_so_far is None
else torch.cat((output_so_far, selected_pert_gen_tok_text), dim=1)
)
if(output_so_far.shape[1]>50):
index=(output_so_far.squeeze(dim=0)==50256).nonzero()[-3].cpu().numpy()[0]
output_so_far=output_so_far[:,index:-1]
#if device == 'cuda':
# torch.cuda.empty_cache()
pert_gen_tok_texts.append(selected_pert_gen_tok_text)
if classifier is not None:
discrim_losses.append(min_discrim_loss)
losses_in_time.append(loss_in_time)
if selected_pert_gen_tok_text is not None:
return tokenizer.decode(selected_pert_gen_tok_text.tolist()[0])
else:
return '|No Response|'
if __name__ == '__main__':
parser = argparse.ArgumentParser()
parser.add_argument(
"--class_label",
"-c",
type=str,
default="neutral",
help="emotion as anger, joy or sadness",
)
parser.add_argument(
"--num_samples",
"-n",
type=int,
default=1,
help="Number of samples to generate the response from",
)
parser.add_argument(
"--seed",
"-s",
type=int,
default=0,
help="a random seed for torch and numpy",
)
args = parser.parse_args()
class_label = args.class_label
num_samples = args.num_samples
seed=args.seed
torch.manual_seed(seed)
np.random.seed(seed)
device = "cuda" if torch.cuda.is_available() else "cpu"
# load pretrained model
model = GPT2LMHeadModel.from_pretrained(pretrained_model, output_hidden_states=True)
# load tokenizer
tokenizer = GPT2Tokenizer.from_pretrained(pretrained_model)
model.to(device)
model.eval()
# Freeze GPT-2 weights
for param in model.parameters():
param.requires_grad = False
with open(discrim_meta_file, 'r') as f:
discrim_meta = json.load(f)
discrim_meta['path'] = discrim_weights
assert discrim_meta['pretrained_model'] == pretrained_model
classifier, class_id = get_classifier(
discrim_meta,
class_label,
device
)
app.run(host='0.0.0.0')