Spaces:
Sleeping
Sleeping
add app.py
Browse files
app.py
ADDED
@@ -0,0 +1,62 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
|
2 |
+
import torch
|
3 |
+
from torch import nn
|
4 |
+
from torchvision import datasets
|
5 |
+
from torchvision.transforms import ToTensor
|
6 |
+
|
7 |
+
|
8 |
+
# Define model
|
9 |
+
class NeuralNetwork(nn.Module):
|
10 |
+
def __init__(self):
|
11 |
+
super().__init__()
|
12 |
+
self.flatten = nn.Flatten()
|
13 |
+
self.linear_relu_stack = nn.Sequential(
|
14 |
+
nn.Linear(28*28, 512),
|
15 |
+
nn.ReLU(),
|
16 |
+
nn.Linear(512, 512),
|
17 |
+
nn.ReLU(),
|
18 |
+
nn.Linear(512, 10)
|
19 |
+
)
|
20 |
+
|
21 |
+
def forward(self, x):
|
22 |
+
x = self.flatten(x)
|
23 |
+
logits = self.linear_relu_stack(x)
|
24 |
+
return logits
|
25 |
+
|
26 |
+
|
27 |
+
model = NeuralNetwork()
|
28 |
+
model.load_state_dict(torch.load("model_mnist_mlp.pth"))
|
29 |
+
|
30 |
+
model.eval()
|
31 |
+
|
32 |
+
import gradio as gr
|
33 |
+
from torchvision import transforms
|
34 |
+
|
35 |
+
def predict(image):
|
36 |
+
tsr_image = transforms.ToTensor()(image)
|
37 |
+
|
38 |
+
with torch.no_grad():
|
39 |
+
pred = model(tsr_image)
|
40 |
+
prob = torch.nn.functional.softmax(pred[0], dim=0)
|
41 |
+
|
42 |
+
confidences = {i: float(prob[i]) for i in range(10)}
|
43 |
+
|
44 |
+
return confidences
|
45 |
+
|
46 |
+
|
47 |
+
with gr.Blocks(css=".gradio-container {background:lightyellow;color:red;}", title="テスト"
|
48 |
+
) as demo:
|
49 |
+
gr.HTML('<div style="font-size:12pt; text-align:center; color:yellow;"MNIST 分類器</div>')
|
50 |
+
|
51 |
+
with gr.Row():
|
52 |
+
input_image = gr.Image(label="画像入力", type="pil", image_mode="L", shape=(28, 28), invert_colors=True)
|
53 |
+
|
54 |
+
output_label=gr.Label(label="予測確率", num_top_classes=5)
|
55 |
+
|
56 |
+
send_btn = gr.Button("予測する")
|
57 |
+
send_btn.click(fn=predict, inputs=input_image, outputs=output_label)
|
58 |
+
|
59 |
+
# demo.queue(concurrency_count=3)
|
60 |
+
demo.launch()
|
61 |
+
|
62 |
+
### EOF ###
|