Spaces:
Runtime error
Runtime error
Commit
•
c864490
1
Parent(s):
73e1547
Initial version of the app
Browse files- app.py +137 -0
- requirements.txt +2 -0
app.py
ADDED
@@ -0,0 +1,137 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import os
|
2 |
+
import re
|
3 |
+
import gradio as gr
|
4 |
+
import torch
|
5 |
+
from huggingface_hub import hf_hub_download
|
6 |
+
from torch.utils.data import DataLoader, Dataset
|
7 |
+
|
8 |
+
|
9 |
+
|
10 |
+
|
11 |
+
model_file_path = hf_hub_download("meliascosta/wiki_hiagm", "hiagm_dyn_quant.p",
|
12 |
+
use_auth_token=os.environ['TOKEN'])
|
13 |
+
vocab_file_path = hf_hub_download("meliascosta/wiki_hiagm", "vocab/word.dict",
|
14 |
+
use_auth_token=os.environ['TOKEN'])
|
15 |
+
label_file_path = hf_hub_download("meliascosta/wiki_hiagm", "vocab/label.dict",
|
16 |
+
use_auth_token=os.environ['TOKEN'])
|
17 |
+
|
18 |
+
# Load model
|
19 |
+
# m = torch.load('./temp_dyn.p')
|
20 |
+
model = torch.load(model_file_path, map_location=torch.device('cpu'))
|
21 |
+
|
22 |
+
model.eval()
|
23 |
+
|
24 |
+
MAX_INPUT_LENGTH = 256
|
25 |
+
K_TOP = 3
|
26 |
+
BATCH_SIZE = 512
|
27 |
+
P_THRESHOLD = 0.5
|
28 |
+
|
29 |
+
class CustomImageDataset(Dataset):
|
30 |
+
def __init__(self, samples):
|
31 |
+
self.samples = samples
|
32 |
+
|
33 |
+
def __len__(self):
|
34 |
+
return len(self.samples)
|
35 |
+
|
36 |
+
def __getitem__(self, idx):
|
37 |
+
return self.samples[idx]
|
38 |
+
|
39 |
+
|
40 |
+
# Load vocab mappings
|
41 |
+
|
42 |
+
v2i = {}
|
43 |
+
i2v = {}
|
44 |
+
with open(vocab_file_path) as f:
|
45 |
+
for i, line in enumerate(f):
|
46 |
+
data = line.rstrip().split('\t')
|
47 |
+
assert len(data) == 2
|
48 |
+
v2i[data[0]] = i
|
49 |
+
i2v[i] = data[0]
|
50 |
+
|
51 |
+
v2i_lab = {}
|
52 |
+
i2v_lab = {}
|
53 |
+
with open(label_file_path) as f:
|
54 |
+
for i, line in enumerate(f):
|
55 |
+
data = line.rstrip().split('\t')
|
56 |
+
assert len(data) == 2
|
57 |
+
v2i_lab[data[0]] = i
|
58 |
+
i2v_lab[i] = data[0]
|
59 |
+
|
60 |
+
|
61 |
+
### PREPROCESSING
|
62 |
+
|
63 |
+
english_stopwords = ['i', 'me', 'my', 'myself', 'we', 'our', 'ours', 'ourselves', 'you', "you're", "you've", "you'll",
|
64 |
+
"you'd", 'your', 'yours', 'yourself', 'yourselves', 'he', 'him', 'his', 'himself', 'she', "she's",
|
65 |
+
'her', 'hers', 'herself', 'it', "it's", 'its', 'itself', 'they', 'them', 'their', 'theirs',
|
66 |
+
'themselves', 'what', 'which', 'who', 'whom', 'this', 'that', "that'll", 'these', 'those', 'am',
|
67 |
+
'is', 'are', 'was', 'were', 'be', 'been', 'being', 'have', 'has', 'had', 'having', 'do', 'does',
|
68 |
+
'did', 'doing', 'a', 'an', 'the', 'and', 'but', 'if', 'or', 'because', 'as', 'until', 'while',
|
69 |
+
'of', 'at', 'by', 'for', 'with', 'about', 'against', 'between', 'into', 'through', 'during',
|
70 |
+
'before', 'after', 'above', 'below', 'to', 'from', 'up', 'down', 'in', 'out', 'on', 'off',
|
71 |
+
'over', 'under', 'again', 'further', 'then', 'once', 'here', 'there', 'when', 'where', 'why',
|
72 |
+
'how', 'all', 'any', 'both', 'each', 'few', 'more', 'most', 'other', 'some', 'such', 'no', 'nor',
|
73 |
+
'not', 'only', 'own', 'same', 'so', 'than', 'too', 'very', 's', 't', 'can', 'will', 'just', 'don',
|
74 |
+
"don't", 'should', "should've", 'now', 'd', 'll', 'm', 'o', 're', 've', 'y', 'ain', 'aren',
|
75 |
+
"aren't", 'couldn', "couldn't", 'didn', "didn't", 'doesn', "doesn't", 'hadn', "hadn't", 'hasn',
|
76 |
+
"hasn't", 'haven', "haven't", 'isn', "isn't", 'ma', 'mightn', "mightn't", 'mustn', "mustn't",
|
77 |
+
'needn', "needn't", 'shan', "shan't", 'shouldn', "shouldn't", 'wasn', "wasn't", 'weren', "weren't",
|
78 |
+
'won', "won't", 'wouldn', "wouldn't", '\\.', '\\?', ',', '\\!', "'s", '']
|
79 |
+
|
80 |
+
|
81 |
+
def clean_stopwords(sample):
|
82 |
+
"""
|
83 |
+
:param sample: List[Str], lower case
|
84 |
+
:return: List[Str]
|
85 |
+
"""
|
86 |
+
return [token for token in sample if token not in english_stopwords]
|
87 |
+
|
88 |
+
|
89 |
+
def clean_str(string):
|
90 |
+
"""
|
91 |
+
Original Source: https://github.com/yoonkim/CNN_sentence/blob/master/process_data.py
|
92 |
+
:param string: Str
|
93 |
+
:return -> Str
|
94 |
+
"""
|
95 |
+
string = string.strip().strip('"')
|
96 |
+
string = re.sub(r"[^A-Za-z(),!?\.\'\`]", " ", string)
|
97 |
+
string = re.sub(r"\'s", " \'s", string)
|
98 |
+
string = re.sub(r"\'ve", " \'ve", string)
|
99 |
+
string = re.sub(r"n\'t", " n\'t", string)
|
100 |
+
string = re.sub(r"\'re", " \'re", string)
|
101 |
+
string = re.sub(r"\'d", " \'d", string)
|
102 |
+
string = re.sub(r"\'ll", " \'ll", string)
|
103 |
+
string = re.sub(r",", " , ", string)
|
104 |
+
string = re.sub(r"\.", " \. ", string)
|
105 |
+
string = re.sub(r"\"", " , ", string)
|
106 |
+
string = re.sub(r"!", " ! ", string)
|
107 |
+
string = re.sub(r"\(", " \( ", string)
|
108 |
+
string = re.sub(r"\)", " \) ", string)
|
109 |
+
string = re.sub(r"\?", " \? ", string)
|
110 |
+
string = re.sub(r"\s{2,}", " ", string)
|
111 |
+
return string.strip().lower()
|
112 |
+
|
113 |
+
|
114 |
+
def preprocess_line(sample):
|
115 |
+
"""
|
116 |
+
:param sample: Str, "The sample would be tokenized and filtered according to the stopwords list"
|
117 |
+
:return: token_list -> List[Str]
|
118 |
+
"""
|
119 |
+
sample = clean_str(sample.lstrip().rstrip())
|
120 |
+
token_list = clean_stopwords(sample.split(' '))
|
121 |
+
return {'token': token_list, 'label': []}
|
122 |
+
|
123 |
+
|
124 |
+
|
125 |
+
def predict(line):
|
126 |
+
preprocessed_line = preprocess_line(line)
|
127 |
+
sample = {}
|
128 |
+
sample["token"] = [v2i.get(v.lower(), v2i['<OOV>']) for v in preprocessed_line['token']]
|
129 |
+
sample["token_len"] = len(sample['token'])
|
130 |
+
sample["token"] = torch.tensor(sample["token"])
|
131 |
+
sample["token_len"] = torch.tensor(sample["token_len"])
|
132 |
+
output_logits = model(next(iter(DataLoader(CustomImageDataset([sample])))))
|
133 |
+
output_probs = torch.sigmoid(output_logits).cpu().tolist()
|
134 |
+
return {i2v_lab[i]:p for i, p in enumerate(output_probs[0]) if p > P_THRESHOLD}
|
135 |
+
|
136 |
+
iface = gr.Interface(fn=predict, inputs="text", outputs=gr.Label(num_top_classes=5))
|
137 |
+
iface.launch()
|
requirements.txt
ADDED
@@ -0,0 +1,2 @@
|
|
|
|
|
|
|
1 |
+
gradio
|
2 |
+
torch
|