Spaces:
Sleeping
Sleeping
Prgckwb
commited on
Commit
·
6c94b18
1
Parent(s):
93013c6
Add app.py
Browse files- app.py +66 -4
- requirements.txt +5 -0
app.py
CHANGED
@@ -1,7 +1,69 @@
|
|
1 |
import gradio as gr
|
|
|
|
|
|
|
2 |
|
3 |
-
|
4 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
5 |
|
6 |
-
|
7 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
import gradio as gr
|
2 |
+
import pandas as pd
|
3 |
+
from gradio.themes import colors
|
4 |
+
from transformers import AutoTokenizer
|
5 |
|
6 |
+
# Function to map tokenized text to IDs
|
7 |
+
def inference(
|
8 |
+
text="",
|
9 |
+
model_id="openai/clip-vit-large-patch14",
|
10 |
+
) -> (list[str, str], pd.DataFrame):
|
11 |
+
if text == "":
|
12 |
+
return [], pd.DataFrame()
|
13 |
+
tokenizer = AutoTokenizer.from_pretrained(model_id)
|
14 |
|
15 |
+
# Use tokenizer to tokenize the text
|
16 |
+
text_inputs = tokenizer(text, return_tensors='pt')
|
17 |
+
|
18 |
+
input_ids = text_inputs['input_ids'].tolist()[0] # Convert tensor to list
|
19 |
+
|
20 |
+
# Create pairs of tokens and IDs
|
21 |
+
tokens = [tokenizer.decode([id_]) for id_ in input_ids]
|
22 |
+
token_pairs = []
|
23 |
+
|
24 |
+
for token, id_ in zip(tokens, input_ids):
|
25 |
+
token_pairs.append((token, str(id_)))
|
26 |
+
|
27 |
+
# Count the number of characters and tokens
|
28 |
+
pos_count = pd.DataFrame({
|
29 |
+
"Char Count": [len(text)],
|
30 |
+
"Token Count": [len(token_pairs)]
|
31 |
+
})
|
32 |
+
return token_pairs, pos_count
|
33 |
+
|
34 |
+
|
35 |
+
if __name__ == '__main__':
|
36 |
+
iface = gr.Interface(
|
37 |
+
fn=inference,
|
38 |
+
inputs=[
|
39 |
+
gr.Textbox(label="Text"),
|
40 |
+
gr.Dropdown(
|
41 |
+
label="Model",
|
42 |
+
choices=[
|
43 |
+
"openai/clip-vit-large-patch14",
|
44 |
+
"google-bert/bert-base-uncased",
|
45 |
+
"google/flan-t5-base",
|
46 |
+
"openai-community/gpt2",
|
47 |
+
],
|
48 |
+
value="openai/clip-vit-large-patch14"
|
49 |
+
),
|
50 |
+
],
|
51 |
+
outputs=[
|
52 |
+
gr.Highlightedtext(label="Highlighted Text"),
|
53 |
+
gr.Dataframe(label="Position Count"),
|
54 |
+
],
|
55 |
+
examples=[
|
56 |
+
["When I told my computer I needed a break, it froze.", "openai/clip-vit-large-patch14"],
|
57 |
+
["Yesterday, I thought my cat was studying for her degree in philosophy because she sat on my book, "
|
58 |
+
"but turns out she was just trying to hatch a plot to steal my dinner.", "openai/clip-vit-large-patch14"],
|
59 |
+
["The square root of x is the cube root of y. What is y to the power of 2, if x = 4?",
|
60 |
+
"google/flan-t5-base"]
|
61 |
+
],
|
62 |
+
cache_examples=True,
|
63 |
+
title="TokenVisor",
|
64 |
+
description="Visualize how the Tokenizer used in Hugging Face's Transformers library tokenizes text.",
|
65 |
+
theme=gr.Theme(primary_hue=colors.green, secondary_hue=colors.yellow),
|
66 |
+
allow_flagging="never",
|
67 |
+
|
68 |
+
)
|
69 |
+
iface.launch()
|
requirements.txt
ADDED
@@ -0,0 +1,5 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
torch
|
2 |
+
transformers
|
3 |
+
safetensors
|
4 |
+
accelerate
|
5 |
+
diffusers
|