File size: 3,256 Bytes
75f1b92
7af13f4
da4611f
48193db
7af13f4
 
 
 
75f1b92
 
 
 
7af13f4
 
75f1b92
 
 
7af13f4
 
48193db
7af13f4
 
 
75f1b92
 
 
7af13f4
 
75f1b92
7af13f4
 
 
75f1b92
 
 
7af13f4
75f1b92
da4611f
75f1b92
7af13f4
75f1b92
 
 
 
 
 
 
7af13f4
 
 
 
 
 
 
 
 
 
 
 
 
 
 
75f1b92
19964c7
75f1b92
 
19964c7
 
 
 
 
 
7af13f4
 
 
 
d3d7595
7af13f4
06b445f
a14bbf4
48193db
75f1b92
 
 
06b445f
75f1b92
 
 
 
06b445f
 
7af13f4
75f1b92
 
 
 
 
 
 
7af13f4
75f1b92
 
 
7af13f4
d3d7595
7af13f4
75f1b92
 
7af13f4
 
 
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
from os import O_ACCMODE
import gradio as gr
import joblib
from transformers import pipeline
import requests.exceptions
from huggingface_hub import HfApi, hf_hub_download
from huggingface_hub.repocard import metadata_load

# work around for error, not happy really
# import os
# os.environ['KMP_DUPLICATE_LIB_OK']='True'

app = gr.Blocks()

model_1 = "juliensimon/distilbert-amazon-shoe-reviews"
model_2 = "juliensimon/distilbert-amazon-shoe-reviews"

def load_agent(model_id_1, model_id_2):
    """
    This function load the agent's results
    """
    # Load the metrics
    metadata_1 = get_metadata(model_id_1)

    # get predictions
    predictions_1 = predict(model_id_1)
    
    # Get the accuracy
    # results_1 = parse_metrics_accuracy(metadata_1)
    
    # Load the metrics
    metadata_2 = get_metadata(model_id_2)

    # get predictions
    predictions_2 = predict(model_id_2)
    # Get the accuracy
    # results_2 = parse_metrics_accuracy(metadata_2)

    return model_id_1, predictions_1, model_id_2, predictions_2

# def parse_metrics_accuracy(meta):
#     if "model-index" not in meta:
#         return None
#     result = meta["model-index"][0]["results"]
#     metrics = result[0]["metrics"]
#     accuracy = metrics[0]["value"]
#     return accuracy

def get_metadata(model_id):
    """
    Get the metadata of the model repo
    :param model_id:
    :return: metadata
    """
    try:
        readme_path = hf_hub_download(model_id, filename="README.md")
        metadata = metadata_load(readme_path)
        print(metadata)
        return metadata
    except requests.exceptions.HTTPError:
        return None

# classifier = pipeline("text-classification", model="juliensimon/distilbert-amazon-shoe-reviews")
                
def predict(review, model_id):
        classifier = pipeline("text-classification", model=model_id)
        prediction = classifier(review)
        print(prediction)
        stars = prediction[0]['label']
        stars = (int)(stars.split('_')[1])+1
        score = 100*prediction[0]['score']
        return "{} {:.0f}%".format("\U00002B50"*stars, score)

with app:
    gr.Markdown(
    """
    # Compare Sentiment Analysis Models 
    
    Type text to predict sentiment.
    """)   
    with gr.Row():
        inp_1= gr.Textbox(label="Type text here.",placeholder="The customer service was satisfactory.")
        out_2 = gr.Textbox(label="Prediction")

    
    # gr.Markdown(
    # """
    # Model Predictions
    # """)
    with gr.Row():
      model1_input = gr.Textbox(label="Model 1")
    with gr.Row():
        btn = gr.Button("Prediction for Model 1")
    btn.click(fn=predict(model_1), inputs=inp_1, outputs=out_2)



    with gr.Row():
      model2_input = gr.Textbox(label="Model 2") 
    with gr.Row():
        btn = gr.Button("Prediction for Model 2")
    btn.click(fn=predict(model_2), inputs=inp_1, outputs=out_2)


    app_button.click(load_agent, inputs=[model1_input, model2_input], outputs=[model1_name, model1_score_output, model2_name, model2_score_output])
    
    # examples = gr.Examples(examples=[["juliensimon/distilbert-amazon-shoe-reviews","juliensimon/distilbert-amazon-shoe-reviews"]],
    #                        inputs=[model1_input, model2_input])

    
app.launch()