fancyfeast
commited on
Commit
•
3be01e3
1
Parent(s):
d240551
Initial commit
Browse files- app.py +79 -0
- requirements.txt +1 -0
app.py
ADDED
@@ -0,0 +1,79 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gr
|
2 |
+
import huggingface_hub
|
3 |
+
from PIL import Image
|
4 |
+
from pathlib import Path
|
5 |
+
import onnxruntime as rt
|
6 |
+
import numpy as np
|
7 |
+
import csv
|
8 |
+
|
9 |
+
|
10 |
+
MODEL_REPO = 'toynya/Z3D-E621-Convnext'
|
11 |
+
THRESHOLD = 0.5
|
12 |
+
DESCRIPTION = """
|
13 |
+
This is a demo of https://huggingface.co/toynya/Z3D-E621-Convnext
|
14 |
+
I am not affiliated with the model author in anyway, this is just a useful tool requested by a user.
|
15 |
+
"""
|
16 |
+
|
17 |
+
|
18 |
+
def prepare_image(image: Image.Image, target_size: int):
|
19 |
+
# Pad image to square
|
20 |
+
image_shape = image.size
|
21 |
+
max_dim = max(image_shape)
|
22 |
+
pad_left = (max_dim - image_shape[0]) // 2
|
23 |
+
pad_top = (max_dim - image_shape[1]) // 2
|
24 |
+
|
25 |
+
padded_image = Image.new("RGB", (max_dim, max_dim), (255, 255, 255))
|
26 |
+
padded_image.paste(image, (pad_left, pad_top))
|
27 |
+
|
28 |
+
# Resize
|
29 |
+
if max_dim != target_size:
|
30 |
+
padded_image = padded_image.resize((target_size, target_size), Image.BICUBIC)
|
31 |
+
|
32 |
+
# Convert to numpy array
|
33 |
+
image_array = np.asarray(padded_image, dtype=np.float32) / 255.0
|
34 |
+
return np.expand_dims(image_array, axis=0)
|
35 |
+
|
36 |
+
|
37 |
+
def predict(image: Image.Image):
|
38 |
+
image_array = prepare_image(image, 448)
|
39 |
+
|
40 |
+
image_array = prepare_image(image, 448)
|
41 |
+
input_name = 'input_1:0'
|
42 |
+
output_name = 'predictions_sigmoid'
|
43 |
+
|
44 |
+
result = session.run([output_name], {input_name: image_array})
|
45 |
+
result = result[0][0]
|
46 |
+
|
47 |
+
scores = {tags[i]: result[i] for i in range(len(result))}
|
48 |
+
predicted_tags = [tag for tag, score in scores.items() if score > THRESHOLD]
|
49 |
+
tag_string = ', '.join(predicted_tags)
|
50 |
+
|
51 |
+
return tag_string, scores
|
52 |
+
|
53 |
+
|
54 |
+
print("Downloading model...")
|
55 |
+
path = huggingface_hub.snapshot_download(MODEL_REPO)
|
56 |
+
print("Loading model...")
|
57 |
+
session = rt.InferenceSession(path / 'model.onnx', providers=["CUDAExecutionProvider", "CPUExecutionProvider"])
|
58 |
+
|
59 |
+
with open(path / 'tags-selected.csv', mode='r', encoding='utf-8') as file:
|
60 |
+
csv_reader = csv.DictReader(file)
|
61 |
+
tags = [row['name'].strip() for row in csv_reader]
|
62 |
+
|
63 |
+
print("Starting server...")
|
64 |
+
|
65 |
+
gradio_app = gr.Interface(
|
66 |
+
predict,
|
67 |
+
inputs=gr.Image(label="Source", sources=['upload', 'webcam'], type='pil'),
|
68 |
+
outputs=[
|
69 |
+
gr.Textbox(label="Tag String"),
|
70 |
+
gr.Label(label="Tag Predictions", num_top_classes=100),
|
71 |
+
],
|
72 |
+
title="JoyTag",
|
73 |
+
description=DESCRIPTION,
|
74 |
+
allow_flagging="never",
|
75 |
+
)
|
76 |
+
|
77 |
+
|
78 |
+
if __name__ == '__main__':
|
79 |
+
gradio_app.launch()
|
requirements.txt
ADDED
@@ -0,0 +1 @@
|
|
|
|
|
1 |
+
numpy==1.26.3
|