-
Notifications
You must be signed in to change notification settings - Fork 0
/
app.py
42 lines (32 loc) · 979 Bytes
/
app.py
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
import torch
import streamlit as st
import pandas as pd
from sentence_transformers import SentenceTransformer, util
st.set_page_config(
page_title="Türkçe Atasözü Önerici",
page_icon="🎈",
)
@st.cache(allow_output_mutation=True)
def load_model():
return SentenceTransformer("dbmdz/bert-base-turkish-uncased")
@st.cache
def load_embeddings():
return torch.load("proverb_embeddings.pt")
@st.cache
def load_data():
return pd.read_csv("atasozleri-vk.csv", index_col=[0], na_filter=False)
model = load_model()
embeddings = load_embeddings()
df = load_data()
st.title("Türkçe Atasözü Önerici")
st.subheader("BERT Tabanlı Atasözü Öneri Sistemi")
girdi = st.text_area(
"Aklından geçenleri buraya dök."
)
if girdi:
girdi_vek = model.encode(girdi)
sim_vec = util.cos_sim(girdi_vek, embeddings)[0]
best_indices = sim_vec.topk(5).indices
results = df[['title', 'anlam']].iloc[best_indices]
st.subheader("Öneriler")
st.write(results)