Dagobert42 commited on
Commit
372cb74
β€’
1 Parent(s): e95aa24

first draft of the space

Browse files
Files changed (4) hide show
  1. README.md +1 -1
  2. app.py +44 -0
  3. requirements.txt +3 -0
  4. sentences.pt +3 -0
README.md CHANGED
@@ -2,7 +2,7 @@
2
  title: Semantic Frame Augmentation
3
  emoji: πŸš€
4
  colorFrom: gray
5
- colorTo: pink
6
  sdk: streamlit
7
  sdk_version: 1.31.1
8
  app_file: app.py
 
2
  title: Semantic Frame Augmentation
3
  emoji: πŸš€
4
  colorFrom: gray
5
+ colorTo: yellow
6
  sdk: streamlit
7
  sdk_version: 1.31.1
8
  app_file: app.py
app.py ADDED
@@ -0,0 +1,44 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ import torch
2
+ import streamlit as st
3
+ from transformers import pipeline
4
+ from random import choice
5
+
6
+ with open("sentences.pt", 'rb') as f:
7
+ sentences = torch.load(f)
8
+
9
+ baseline_classifier = pipeline(
10
+ model="Dagobert42/mobilebert-uncased-biored-finetuned-ner",
11
+ task="ner",
12
+ aggregation_strategy="simple"
13
+ )
14
+ augmented_classifier = pipeline(
15
+ model="Dagobert42/mobilebert-uncased-biored-augmented-ner",
16
+ task="ner",
17
+ aggregation_strategy="simple"
18
+ )
19
+
20
+ st.title("Semantic Frame Augmentation")
21
+ st.caption("Analysing difficult low-resource domains with only a handful of examples")
22
+
23
+ st.write("This space uses a googel/mobilebert-uncased model for named entity ")
24
+ augment = st.toggle('Use augmented model for ', value=False)
25
+
26
+ sentence = choice(sentences)
27
+
28
+ if augment:
29
+ st.write("with augmentation:")
30
+ tokens = augmented_classifier(sentence)
31
+ else:
32
+ st.write("without augmentation:")
33
+ tokens = baseline_classifier(sentence)
34
+
35
+ txt = st.text_area(
36
+ "Text to analyze",
37
+ sentence,
38
+ max_chars=500
39
+ )
40
+
41
+ st.subheader("Entity analysis:")
42
+ for token in tokens:
43
+ st.write(token['entity_group'])
44
+ st.write(sentence[token["start"] : token["end"]])
requirements.txt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ streamlit
2
+ transformers
3
+ torch
sentences.pt ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2081e318c8b942b60a480ce78c4b9938e68be208359a43e27c702f7a6d531f43
3
+ size 178144