Spaces:
Build error
Build error
Delete app.py
Browse files
app.py
DELETED
@@ -1,93 +0,0 @@
|
|
1 |
-
import os
|
2 |
-
import numpy as np
|
3 |
-
import torch
|
4 |
-
import torch.nn.functional as F
|
5 |
-
from components.model import Custom_bert
|
6 |
-
from transformers import AutoTokenizer
|
7 |
-
import gradio as gr
|
8 |
-
|
9 |
-
os.system("gdown https://drive.google.com/uc?id=1whDb0yL_Kqoyx-sIw0sS5xTfb6r_9nlJ")
|
10 |
-
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
|
11 |
-
|
12 |
-
|
13 |
-
def get_batches(input, tokenizer, batch_size=128, max_length=256, device='cpu'):
|
14 |
-
out = tokenizer(input, return_tensors='pt', max_length=max_length, padding='max_length')
|
15 |
-
out['input_ids'], out['attention_mask'] = out['input_ids'].to(device), out['attention_mask'].to(device)
|
16 |
-
input_id_split = torch.split(out['input_ids'], max_length, dim=1)
|
17 |
-
attention_split = torch.split(out['attention_mask'], max_length, dim=1)
|
18 |
-
|
19 |
-
input_id_batches = []
|
20 |
-
attention_batches = []
|
21 |
-
|
22 |
-
i = 0
|
23 |
-
input_length = len(input_id_split)
|
24 |
-
|
25 |
-
while i * batch_size < input_length:
|
26 |
-
if i * batch_size + batch_size <= input_length:
|
27 |
-
input_id_batches.append(list(input_id_split[i * batch_size:(i + 1) * batch_size]))
|
28 |
-
attention_batches.append(list(attention_split[i * batch_size:(i + 1) * batch_size]))
|
29 |
-
else:
|
30 |
-
input_id_batches.append(list(input_id_split[i * batch_size:input_length]))
|
31 |
-
attention_batches.append(list(attention_split[i * batch_size:input_length]))
|
32 |
-
i += 1
|
33 |
-
|
34 |
-
if input_id_batches[-1][-1].shape[1] < max_length:
|
35 |
-
input_id_batches[-1][-1] = F.pad(input_id_batches[-1][-1],
|
36 |
-
(1, max_length - input_id_batches[-1][-1].shape[1] - 1),
|
37 |
-
value=0)
|
38 |
-
attention_batches[-1][-1] = F.pad(attention_batches[-1][-1],
|
39 |
-
(1, max_length - attention_batches[-1][-1].shape[1] - 1),
|
40 |
-
value=1)
|
41 |
-
|
42 |
-
input_id_batches = [torch.cat(batch, dim=0) for batch in input_id_batches]
|
43 |
-
attention_batches = [torch.cat(batch, dim=0) for batch in attention_batches]
|
44 |
-
|
45 |
-
return tuple(zip(input_id_batches, attention_batches))
|
46 |
-
|
47 |
-
|
48 |
-
def predict(input, tokenizer, model, batch_size=128, max_length=256, max_val=-4, min_val=3, score=100):
|
49 |
-
device = model.base.device
|
50 |
-
batches = get_batches(input, tokenizer, batch_size, max_length, device)
|
51 |
-
|
52 |
-
predictions = []
|
53 |
-
|
54 |
-
with torch.no_grad():
|
55 |
-
for input_ids, attention_mask in batches:
|
56 |
-
pred = model(input_ids, attention_mask)
|
57 |
-
pred = score * (pred - min_val) / (max_val - min_val)
|
58 |
-
predictions.append(pred)
|
59 |
-
|
60 |
-
predictions = torch.cat(predictions, dim=0)
|
61 |
-
mean, std = predictions.mean().cpu().item(), predictions.std().cpu().item()
|
62 |
-
mean, std = round(mean, 2), round(std, 2)
|
63 |
-
if np.isnan(std):
|
64 |
-
return f"The reading difficulty score is {mean}."
|
65 |
-
else:
|
66 |
-
return f"""The reading difficulty score is {mean} with a standard deviation of {std}.
|
67 |
-
\nThe 95% confidence interval of the score is {mean - 2 * std} to {mean + 2 * std}."""
|
68 |
-
|
69 |
-
|
70 |
-
if __name__ == "__main__":
|
71 |
-
deberta_loc = "deberta_large_0.pt"
|
72 |
-
deberta_tokenizer = AutoTokenizer.from_pretrained("microsoft/deberta-large", model_max_length=256)
|
73 |
-
|
74 |
-
model = Custom_bert("microsoft/deberta-large")
|
75 |
-
model.load_state_dict(torch.load(deberta_loc))
|
76 |
-
model.eval().to(device)
|
77 |
-
|
78 |
-
|
79 |
-
description = """
|
80 |
-
This tool attempts to estimate how difficult a piece of text is to read by a school child.
|
81 |
-
The underlying model has been developed based on expert ranking of text difficulty for students from grade 3 to 12.
|
82 |
-
The score has been scaled to range from zero (very easy) to one hundred (very difficult).
|
83 |
-
Very long passages will be broken up and reported with the average as well as the standard deviation of the difficulty score.
|
84 |
-
"""
|
85 |
-
|
86 |
-
interface = gr.Interface(fn=lambda x: predict(x, deberta_tokenizer, model, batch_size=4),
|
87 |
-
inputs=gr.inputs.Textbox(lines = 7, label = "Text:",
|
88 |
-
placeholder = "Insert text to be scored here."),
|
89 |
-
outputs='text',
|
90 |
-
title = "Reading Difficulty Analyser",
|
91 |
-
description = description)
|
92 |
-
interface.launch()
|
93 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|