Spaces:
Runtime error
Runtime error
spacy
Browse files- app.py +50 -0
- requirements.txt +2 -0
app.py
ADDED
@@ -0,0 +1,50 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import os
|
2 |
+
|
3 |
+
import gradio as gr
|
4 |
+
from spacy.lang.en import English
|
5 |
+
from transformers import AutoTokenizer
|
6 |
+
|
7 |
+
# download spacy model ---
|
8 |
+
os.system('python -m spacy download en_core_web_sm')
|
9 |
+
|
10 |
+
|
11 |
+
deberta_v3_tokenizer = AutoTokenizer.from_pretrained("microsoft/deberta-v3-base")
|
12 |
+
mistral_tokenizer = AutoTokenizer.from_pretrained("mistralai/Mistral-7B-v0.1")
|
13 |
+
|
14 |
+
en_tokenizer = English().tokenizer
|
15 |
+
|
16 |
+
|
17 |
+
def tokenize_with_spacy(text, tokenizer=en_tokenizer):
|
18 |
+
tokenized_text = tokenizer(text)
|
19 |
+
tokens = [token.text for token in tokenized_text]
|
20 |
+
return tokens
|
21 |
+
|
22 |
+
|
23 |
+
def tokenize_with_hf(text, tokenizer=deberta_v3_tokenizer):
|
24 |
+
tokenized_text = tokenizer.tokenize(text)
|
25 |
+
return tokenized_text
|
26 |
+
|
27 |
+
|
28 |
+
def tokenize(text):
|
29 |
+
s = tokenize_with_spacy(text)
|
30 |
+
d = tokenize_with_hf(text)
|
31 |
+
m = tokenize_with_hf(text, tokenizer=mistral_tokenizer)
|
32 |
+
return s, d, m
|
33 |
+
|
34 |
+
|
35 |
+
with gr.Blocks() as demo:
|
36 |
+
input_text = gr.Textbox(lines=2, placeholder="Input text...")
|
37 |
+
submit_btn = gr.Button(label="Submit")
|
38 |
+
|
39 |
+
spacy_display = gr.JSON(label="Spacy")
|
40 |
+
deb_display = gr.JSON(label="DeBERTa-V3")
|
41 |
+
mistral_display = gr.JSON(label="Mistral")
|
42 |
+
|
43 |
+
# callback ---
|
44 |
+
submit_btn.click(
|
45 |
+
fn=tokenize,
|
46 |
+
outputs=[spacy_display, deb_display, mistral_display],
|
47 |
+
)
|
48 |
+
|
49 |
+
# launch app --------
|
50 |
+
demo.launch()
|
requirements.txt
ADDED
@@ -0,0 +1,2 @@
|
|
|
|
|
|
|
1 |
+
transformers
|
2 |
+
spacy
|