-
Notifications
You must be signed in to change notification settings - Fork 1
/
Copy pathbert-cnn.py
115 lines (86 loc) · 3.53 KB
/
bert-cnn.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
from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
import os
os.environ["TOKENIZERS_PARALLELISM"] = "false"
def distilbert_model(storyfile):
tokenizer = AutoTokenizer.from_pretrained("sshleifer/distilbart-cnn-12-6")
model = AutoModelForSeq2SeqLM.from_pretrained("sshleifer/distilbart-cnn-12-6")
# storyPath = "stories/" + storyfile
storyPath = storyfile
f = open(storyPath, "r")
STORY = f.read()
# inputs = tokenizer("summarize: " + STORY, return_tensors="pt", max_length=10000, truncation=True)
inputs = tokenizer(STORY, return_tensors="pt", truncation=True)
outputs = model.generate(
inputs["input_ids"], max_length=400, min_length=40, length_penalty=2.0, num_beams=4, early_stopping=False
)
summary = tokenizer.decode(outputs[0])
summary = summary.strip('</s>')
summary = summary.strip('<s>')
summary += '\n\n\n'
sf = storyfile.strip(".txt")
sf = sf.split("_")
summaryName = 'summary' + '_' + sf[1]
summaryFilename = 'summaries/' + summaryName + '.txt'
f = open(summaryFilename, "a")
f.write(summary)
f.close()
return summary
def googleT5_model(storyfile):
model = AutoModelForSeq2SeqLM.from_pretrained("t5-base")
tokenizer = AutoTokenizer.from_pretrained("t5-base")
# storyPath = "stories/" + storyfile
storyPath = storyfile
f = open(storyPath, "r")
STORY = f.read()
# T5 uses a max_length of 512 so we cut the article to 512 tokens.
inputs = tokenizer("summarize: " + STORY, return_tensors="pt", max_length=10000, truncation=True)
outputs = model.generate(
inputs["input_ids"], max_length=400, min_length=40, length_penalty=2.0, num_beams=4, early_stopping=False
)
summary = tokenizer.decode(outputs[0])
summary = summary.strip('<pad>')
summary = summary.strip('</s>')
summary += '\n\n\n'
sf = storyfile.strip(".txt")
sf = sf.split("_")
summaryName = 'summary' + '_' + sf[1]
summaryFilename = 'summaries/' + summaryName + '.txt'
f = open(summaryFilename, "a")
f.write(summary)
f.close()
return summary
def bert_model(storyfile):
tokenizer = AutoTokenizer.from_pretrained("philschmid/bart-large-cnn-samsum")
model = AutoModelForSeq2SeqLM.from_pretrained("philschmid/bart-large-cnn-samsum")
# storyPath = "stories/" + storyfile
storyPath = storyfile
f = open(storyPath, "r")
STORY = f.read()
# inputs = tokenizer("summarize: " + STORY, return_tensors="pt", max_length=100000, truncation=True)
inputs = tokenizer(STORY, return_tensors="pt", truncation=True)
outputs = model.generate(
inputs["input_ids"], max_length=400, min_length=40, length_penalty=2.0, num_beams=4, early_stopping=False
)
summary = tokenizer.decode(outputs[0])
summary = summary.strip('</s>')
summary = summary.strip('<s>')
summary += '\n\n\n'
sf = storyfile.strip(".txt")
sf = sf.split("_")
summaryName = 'summary' + '_' + sf[1]
summaryFilename = 'summaries/' + summaryName + '.txt'
f = open(summaryFilename, "a")
f.write(summary)
f.close()
return summaryFilename
# for i in range(0, 10):
# STORY_NAME = 'story_' + str(i) + '.txt'
# summary_bert = bert_model(STORY_NAME)
# print('Bert-cnn generated summary: \n')
# print(summary_bert)
# summary_googleT5 = googleT5_model(STORY_NAME)
# print('google T5 generated summary: \n')
# print(summary_googleT5)
# summary_distilbert = distilbert_model(STORY_NAME)
# print('distilbert generated summary: \n')
# print(summary_distilbert)