Spaces:
Paused
Paused
Commit
·
9a229a7
1
Parent(s):
7d3498d
Update app.py
Browse files
app.py
CHANGED
@@ -1,10 +1,65 @@
|
|
1 |
import gradio as gr
|
2 |
-
|
3 |
|
4 |
-
|
5 |
|
6 |
-
def
|
7 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
8 |
|
9 |
iface = gr.Interface(fn=greet, inputs="text", outputs="text")
|
10 |
iface.launch()
|
|
|
1 |
import gradio as gr
|
2 |
+
from transformers import BertForQuestionAnswering
|
3 |
|
4 |
+
model = BertForQuestionAnswering.from_pretrained("bert-base-uncased")
|
5 |
|
6 |
+
def get_prediction(context, question):
|
7 |
+
inputs = tokenizer.encode_plus(question, context, return_tensors='pt').to(device)
|
8 |
+
outputs = model(**inputs)
|
9 |
+
|
10 |
+
answer_start = torch.argmax(outputs[0])
|
11 |
+
answer_end = torch.argmax(outputs[1]) + 1
|
12 |
+
|
13 |
+
answer = tokenizer.convert_tokens_to_string(tokenizer.convert_ids_to_tokens(inputs['input_ids'][0][answer_start:answer_end]))
|
14 |
+
|
15 |
+
return answer
|
16 |
+
|
17 |
+
def normalize_text(s):
|
18 |
+
"""Removing articles and punctuation, and standardizing whitespace are all typical text processing steps."""
|
19 |
+
import string, re
|
20 |
+
def remove_articles(text):
|
21 |
+
regex = re.compile(r"\b(a|an|the)\b", re.UNICODE)
|
22 |
+
return re.sub(regex, " ", text)
|
23 |
+
def white_space_fix(text):
|
24 |
+
return " ".join(text.split())
|
25 |
+
def remove_punc(text):
|
26 |
+
exclude = set(string.punctuation)
|
27 |
+
return "".join(ch for ch in text if ch not in exclude)
|
28 |
+
def lower(text):
|
29 |
+
return text.lower()
|
30 |
+
|
31 |
+
return white_space_fix(remove_articles(remove_punc(lower(s))))
|
32 |
+
|
33 |
+
def exact_match(prediction, truth):
|
34 |
+
return bool(normalize_text(prediction) == normalize_text(truth))
|
35 |
+
|
36 |
+
def compute_f1(prediction, truth):
|
37 |
+
pred_tokens = normalize_text(prediction).split()
|
38 |
+
truth_tokens = normalize_text(truth).split()
|
39 |
+
|
40 |
+
# if either the prediction or the truth is no-answer then f1 = 1 if they agree, 0 otherwise
|
41 |
+
if len(pred_tokens) == 0 or len(truth_tokens) == 0:
|
42 |
+
return int(pred_tokens == truth_tokens)
|
43 |
+
|
44 |
+
common_tokens = set(pred_tokens) & set(truth_tokens)
|
45 |
+
|
46 |
+
# if there are no common tokens then f1 = 0
|
47 |
+
if len(common_tokens) == 0:
|
48 |
+
return 0
|
49 |
+
|
50 |
+
prec = len(common_tokens) / len(pred_tokens)
|
51 |
+
rec = len(common_tokens) / len(truth_tokens)
|
52 |
+
|
53 |
+
return round(2 * (prec * rec) / (prec + rec), 2)
|
54 |
+
|
55 |
+
def question_answer(context, question):
|
56 |
+
prediction = get_prediction(context,question)
|
57 |
+
return prediction
|
58 |
+
|
59 |
+
def greet(texts):
|
60 |
+
# for question, answer in texts:
|
61 |
+
# question_answer(context, question)
|
62 |
+
return texts
|
63 |
|
64 |
iface = gr.Interface(fn=greet, inputs="text", outputs="text")
|
65 |
iface.launch()
|