Spaces:
Sleeping
Sleeping
dbleek
commited on
Commit
·
d2e0837
1
Parent(s):
930a750
implemented patent classifier
Browse files
app_pt.py
ADDED
@@ -0,0 +1,59 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import streamlit as st
|
2 |
+
import torch
|
3 |
+
from datasets import load_dataset
|
4 |
+
from transformers import AutoTokenizer
|
5 |
+
from transformers import AutoModelForSequenceClassification
|
6 |
+
from transformers import pipeline
|
7 |
+
|
8 |
+
dataset_dict = load_dataset('HUPD/hupd',
|
9 |
+
name='sample',
|
10 |
+
data_files="https://huggingface.co/datasets/HUPD/hupd/blob/main/hupd_metadata_2022-02-22.feather",
|
11 |
+
icpr_label=None,
|
12 |
+
train_filing_start_date='2016-01-01',
|
13 |
+
train_filing_end_date='2016-01-21',
|
14 |
+
val_filing_start_date='2016-01-22',
|
15 |
+
val_filing_end_date='2016-01-31',
|
16 |
+
)
|
17 |
+
model = torch.load("/workspaces/cs-gy-6613-project/patent_classification(1).pt", map_location=torch.device('cpu'))
|
18 |
+
tokenizer = AutoTokenizer.from_pretrained("distilbert-base-uncased")
|
19 |
+
classifier = pipeline("text-classification", model=model, tokenizer=tokenizer)
|
20 |
+
filtered_dataset = dataset_dict['validation'].filter(lambda e: e['decision'] == 'ACCEPTED' or e['decision'] == 'REJECTED')
|
21 |
+
dataset = filtered_dataset.shuffle(seed=42).select(range(20))
|
22 |
+
dataset = dataset.sort("patent_number")
|
23 |
+
applications = {}
|
24 |
+
|
25 |
+
for ds_index, example in enumerate(dataset):
|
26 |
+
applications.update({example['patent_number']: ds_index })
|
27 |
+
|
28 |
+
def load_patent():
|
29 |
+
selected_application = dataset.select([applications[st.session_state.id]])
|
30 |
+
st.session_state.abstract = selected_application['abstract'][0]
|
31 |
+
st.session_state.claims = selected_application['claims'][0]
|
32 |
+
st.session_state.title = selected_application['title'][0]
|
33 |
+
|
34 |
+
st.title("CS-GY-6613 Project Milestone 3")
|
35 |
+
|
36 |
+
patent_number = st.selectbox("Select a patent application:", applications, on_change=load_patent, key="id")
|
37 |
+
title = st.text_area("Title", key="title", value=dataset[0]['title'], height=50)
|
38 |
+
|
39 |
+
with st.form('Details'):
|
40 |
+
abstract = st.text_area("Abstract", key="abstract", value=dataset[0]['abstract'], height=200)
|
41 |
+
claims = st.text_area("Claims", key="claims", value=dataset[0]['abstract'], height=200)
|
42 |
+
submitted = st.form_submit_button("Get Patentability Score")
|
43 |
+
|
44 |
+
if submitted:
|
45 |
+
selected_application = dataset.select([applications[st.session_state.id]])
|
46 |
+
res = classifier(abstract, claims)
|
47 |
+
if res[0]["label"] == 'LABEL_0':
|
48 |
+
pred = "ACCEPTED"
|
49 |
+
elif res[0]["label"] == 'LABEL_1':
|
50 |
+
pred = "REJECTED"
|
51 |
+
score = res[0]["score"]
|
52 |
+
label = selected_application['decision'][0]
|
53 |
+
result = st.markdown("This text was classified as **{}** with a confidence score of **{}**.".format(pred, score))
|
54 |
+
check = st.markdown("Actual Label: **{}**.".format(label))
|
55 |
+
|
56 |
+
|
57 |
+
|
58 |
+
|
59 |
+
|