File size: 3,431 Bytes
dbb9b6d
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
from transformers import AutoTokenizer
from transformers import AutoModelForSequenceClassification
import torch
from torch.nn import functional as F
import numpy as np
import json
from utils.client import generate_seo_metatitle



# id2label= {0: 'Commercial',
#  1: 'Informational',
#  2: 'Navigational',
#  3: 'Local',
#  4: 'Transactional'}

# label2id= {'Commercial': 0,
#  'Informational': 1,
#  'Navigational': 2,
#  'Local': 3,
#  'Transactional': 4}


# removed local
id2label= {0: 'Commercial', 1: 'Informational', 2: 'Navigational', 3: 'Transactional'}
label2id= {'Commercial': 0, 'Informational': 1, 'Navigational': 2, 'Transactional': 3}

model_name= "intent_classification_model_with_metatitle_with_local2/checkpoint-2700"
tokenizer = AutoTokenizer.from_pretrained(model_name)

model = AutoModelForSequenceClassification.from_pretrained(model_name).to("cuda")


# probabilities = 1 / (1 + np.exp(-logit_score))
def logit2prob(logit):
    # odds =np.exp(logit)
    # prob = odds / (1 + odds)
    prob= 1/(1+ np.exp(-logit))
    return np.round(prob, 3)





def get_intent_one_by_one(keyword:str):
    inputs = tokenizer(generate_seo_metatitle(keyword), padding=True, truncation=True, return_tensors="pt").to("cuda")
    with torch.no_grad():
        logits = model(**inputs).logits
        
    # print("logits: ", logits)
    # predicted_class_id = logits.argmax().item()
    
    # get probabilities using softmax from logit score and convert it to numpy array
    # probabilities_scores = F.softmax(logits.cpu(), dim = -1).numpy()[0]
    individual_probabilities_scores = logit2prob(logits.cpu().numpy()[0])
    
    score_list= []
    
    for i in range(len(id2label)):
        label= id2label[i]
        
        score= individual_probabilities_scores[i]
        if score>0.5:
            score_list.append(
                        (label, score)
                    )
        # if score>=0.5: 
        #     score_list.append(
        #         (id2label[i], score)
        #     )
            
    if len(score_list)==0:
        score_list.append(("undefined",1))        
    score_list.sort(
        key= lambda x: x[1], reverse=True
    )
            
    return score_list
    # return (np.argmax(i), id2label[np.argmax(i)])
    



def get_intent_one_by_one_test(metatitle:str):
    inputs = tokenizer(metatitle,padding=True, truncation=True, return_tensors="pt").to("cuda")
    with torch.no_grad():
        logits = model(**inputs).logits
        
    # print("logits: ", logits)
    # predicted_class_id = logits.argmax().item()
    
    # get probabilities using softmax from logit score and convert it to numpy array
    # probabilities_scores = F.softmax(logits.cpu(), dim = -1).numpy()[0]
    individual_probabilities_scores = logit2prob(logits.cpu().numpy()[0])
    
    score_list= []
    
    for i in range(len(id2label)):
        label= id2label[i]
        
        score= individual_probabilities_scores[i]
        if score>0.5:
            score_list.append(
                        (label, score)
                    )
        # if score>=0.5: 
        #     score_list.append(
        #         (id2label[i], score)
        #     )
            
    if len(score_list)==0:
        score_list.append(("undefined",1))        
    score_list.sort(
        key= lambda x: x[1], reverse=True
    )
            
    return score_list
    # return (np.argmax(i), id2label[np.argmax(i)])