Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -1,12 +1,104 @@
|
|
1 |
-
import gradio as gr
|
2 |
-
|
3 |
-
|
4 |
-
|
5 |
-
|
6 |
-
|
7 |
-
|
8 |
-
|
9 |
-
|
10 |
-
|
11 |
-
|
12 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gr
|
2 |
+
import torch
|
3 |
+
import torch.nn as nn
|
4 |
+
from torchvision import transforms, models
|
5 |
+
import pickle
|
6 |
+
|
7 |
+
|
8 |
+
with open('class_names.pkl', 'rb') as f:
|
9 |
+
class_names = pickle.load(f)
|
10 |
+
|
11 |
+
# 加载训练好的模型
|
12 |
+
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
|
13 |
+
|
14 |
+
model = models.resnet50(weights=None)
|
15 |
+
model.fc = nn.Sequential(
|
16 |
+
nn.Dropout(0.2),
|
17 |
+
nn.Linear(model.fc.in_features, len(class_names))
|
18 |
+
)
|
19 |
+
|
20 |
+
# 加载模型权重
|
21 |
+
model.load_state_dict(torch.load('best_model.pth', map_location=device, weights_only=True))
|
22 |
+
model = model.to(device)
|
23 |
+
model.eval()
|
24 |
+
|
25 |
+
# 定义与训练时相同的预处理流程
|
26 |
+
preprocess = transforms.Compose([
|
27 |
+
transforms.Resize((100, 100)),
|
28 |
+
transforms.ToTensor(),
|
29 |
+
transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
|
30 |
+
])
|
31 |
+
|
32 |
+
|
33 |
+
def predict_image(img):
|
34 |
+
img = img.convert('RGB')
|
35 |
+
|
36 |
+
# 应用预处理
|
37 |
+
input_tensor = preprocess(img)
|
38 |
+
|
39 |
+
# 添加批次维度并移动到设备
|
40 |
+
input_batch = input_tensor.unsqueeze(0).to(device)
|
41 |
+
|
42 |
+
# 预测
|
43 |
+
with torch.no_grad():
|
44 |
+
output = model(input_batch)
|
45 |
+
|
46 |
+
# 计算概率
|
47 |
+
probabilities = torch.nn.functional.softmax(output[0], dim=0)
|
48 |
+
|
49 |
+
# 获取前3个预测结果
|
50 |
+
top3_probs, top3_indices = torch.topk(probabilities, 3)
|
51 |
+
|
52 |
+
results = {
|
53 |
+
class_names[i]: p.item()
|
54 |
+
for p, i in zip(top3_probs, top3_indices)
|
55 |
+
}
|
56 |
+
|
57 |
+
# 获取最佳预测结果
|
58 |
+
best_class = class_names[top3_indices[0]]
|
59 |
+
best_conf = top3_probs[0].item() * 100
|
60 |
+
|
61 |
+
# 保存结果
|
62 |
+
with open('prediction_results.txt', 'a') as f:
|
63 |
+
f.write(f"Image: {img}\n"
|
64 |
+
f"Predicted: {best_class}\n"
|
65 |
+
f"Confidence: {best_conf:.2f}%\n"
|
66 |
+
f"Top 3: {results}\n"
|
67 |
+
f"------------------------\n")
|
68 |
+
|
69 |
+
return best_class, best_conf, results
|
70 |
+
|
71 |
+
# 创建Gradio界面
|
72 |
+
def create_interface():
|
73 |
+
examples = [
|
74 |
+
"data/r0_0_100.jpg",
|
75 |
+
"data/r0_18_100.jpg"
|
76 |
+
]
|
77 |
+
|
78 |
+
with gr.Blocks(title="Fruit Classification", theme=gr.themes.Soft()) as demo:
|
79 |
+
gr.Markdown("# 🍎 水果识别系统")
|
80 |
+
|
81 |
+
with gr.Row():
|
82 |
+
with gr.Column():
|
83 |
+
image_input = gr.Image(type="pil", label="上传图像")
|
84 |
+
gr.Examples(examples=examples, inputs=image_input)
|
85 |
+
submit_btn = gr.Button("分类", variant="primary")
|
86 |
+
|
87 |
+
with gr.Column():
|
88 |
+
best_pred = gr.Textbox(label="预测结果")
|
89 |
+
confidence = gr.Textbox(label="置信度")
|
90 |
+
full_results = gr.Label(label="Top 3", num_top_classes=3)
|
91 |
+
|
92 |
+
# ‘分类’按钮点击事件
|
93 |
+
submit_btn.click(
|
94 |
+
fn=predict_image,
|
95 |
+
inputs=image_input,
|
96 |
+
outputs=[best_pred, confidence, full_results]
|
97 |
+
)
|
98 |
+
|
99 |
+
return demo
|
100 |
+
|
101 |
+
|
102 |
+
if __name__ == "__main__":
|
103 |
+
interface = create_interface()
|
104 |
+
interface.launch(share=True)
|