Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
@@ -5,10 +5,28 @@ from torchvision import transforms, models
|
|
5 |
import pickle
|
6 |
from resnest.torch import resnest50
|
7 |
|
|
|
8 |
with open('class_names.pkl', 'rb') as f:
|
9 |
-
|
10 |
-
|
11 |
-
#
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
12 |
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
|
13 |
|
14 |
model = resnest50(pretrained=None)
|
@@ -17,12 +35,12 @@ model.fc = nn.Sequential(
|
|
17 |
nn.Linear(model.fc.in_features, len(class_names))
|
18 |
)
|
19 |
|
20 |
-
#
|
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 |
-
#
|
26 |
preprocess = transforms.Compose([
|
27 |
transforms.Resize((100, 100)),
|
28 |
transforms.ToTensor(),
|
@@ -33,20 +51,20 @@ preprocess = transforms.Compose([
|
|
33 |
def predict_image(img):
|
34 |
img = img.convert('RGB')
|
35 |
|
36 |
-
#
|
37 |
input_tensor = preprocess(img)
|
38 |
|
39 |
-
#
|
40 |
input_batch = input_tensor.unsqueeze(0).to(device)
|
41 |
|
42 |
-
#
|
43 |
with torch.no_grad():
|
44 |
output = model(input_batch)
|
45 |
|
46 |
-
#
|
47 |
probabilities = torch.nn.functional.softmax(output[0], dim=0)
|
48 |
|
49 |
-
#
|
50 |
top3_probs, top3_indices = torch.topk(probabilities, 3)
|
51 |
|
52 |
results = {
|
@@ -54,42 +72,42 @@ def predict_image(img):
|
|
54 |
for p, i in zip(top3_probs, top3_indices)
|
55 |
}
|
56 |
|
57 |
-
#
|
58 |
best_class = class_names[top3_indices[0]]
|
59 |
best_conf = top3_probs[0].item() * 100
|
60 |
|
61 |
-
#
|
62 |
with open('/tmp/prediction_results.txt', 'a') as f:
|
63 |
-
f.write(f"
|
64 |
-
f"
|
65 |
-
f"
|
66 |
f"Top 3: {results}\n"
|
67 |
f"------------------------\n")
|
68 |
|
69 |
-
return best_class, best_conf, results
|
70 |
|
71 |
-
#
|
72 |
def create_interface():
|
73 |
examples = [
|
74 |
"r0_0_100.jpg",
|
75 |
"r0_18_100.jpg"
|
76 |
]
|
77 |
|
78 |
-
with gr.Blocks(title="
|
79 |
-
gr.Markdown("# 🍎
|
80 |
|
81 |
with gr.Row():
|
82 |
with gr.Column():
|
83 |
-
image_input = gr.Image(type="pil", label="
|
84 |
gr.Examples(examples=examples, inputs=image_input)
|
85 |
-
submit_btn = gr.Button("
|
86 |
|
87 |
with gr.Column():
|
88 |
-
best_pred = gr.Textbox(label="
|
89 |
-
confidence = gr.Textbox(label="
|
90 |
full_results = gr.Label(label="Top 3", num_top_classes=3)
|
91 |
|
92 |
-
#
|
93 |
submit_btn.click(
|
94 |
fn=predict_image,
|
95 |
inputs=image_input,
|
@@ -101,4 +119,4 @@ def create_interface():
|
|
101 |
|
102 |
if __name__ == "__main__":
|
103 |
interface = create_interface()
|
104 |
-
interface.launch(share=False)
|
|
|
5 |
import pickle
|
6 |
from resnest.torch import resnest50
|
7 |
|
8 |
+
# Carregar nomes das classes e criar mapeamento para português
|
9 |
with open('class_names.pkl', 'rb') as f:
|
10 |
+
class_names_en = pickle.load(f)
|
11 |
+
|
12 |
+
# Dicionário de tradução das classes para português
|
13 |
+
class_names_pt = {
|
14 |
+
'apple': 'maçã',
|
15 |
+
'banana': 'banana',
|
16 |
+
'cherry': 'cereja',
|
17 |
+
'chico': 'sapoti', # nome em português para chico/fruit
|
18 |
+
'grape': 'uva',
|
19 |
+
'kiwi': 'kiwi',
|
20 |
+
'mango': 'manga',
|
21 |
+
'orange': 'laranja',
|
22 |
+
'pear': 'pera',
|
23 |
+
'tomato': 'tomate'
|
24 |
+
}
|
25 |
+
|
26 |
+
# Criar lista de nomes em português na mesma ordem que class_names_en
|
27 |
+
class_names = [class_names_pt[en] for en in class_names_en]
|
28 |
+
|
29 |
+
# Carregar o modelo treinado
|
30 |
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
|
31 |
|
32 |
model = resnest50(pretrained=None)
|
|
|
35 |
nn.Linear(model.fc.in_features, len(class_names))
|
36 |
)
|
37 |
|
38 |
+
# Carregar os pesos do modelo
|
39 |
model.load_state_dict(torch.load('best_model.pth', map_location=device, weights_only=True))
|
40 |
model = model.to(device)
|
41 |
model.eval()
|
42 |
|
43 |
+
# Definir o mesmo pré-processamento usado no treinamento
|
44 |
preprocess = transforms.Compose([
|
45 |
transforms.Resize((100, 100)),
|
46 |
transforms.ToTensor(),
|
|
|
51 |
def predict_image(img):
|
52 |
img = img.convert('RGB')
|
53 |
|
54 |
+
# Aplicar pré-processamento
|
55 |
input_tensor = preprocess(img)
|
56 |
|
57 |
+
# Adicionar dimensão de batch e mover para o dispositivo
|
58 |
input_batch = input_tensor.unsqueeze(0).to(device)
|
59 |
|
60 |
+
# Fazer previsão
|
61 |
with torch.no_grad():
|
62 |
output = model(input_batch)
|
63 |
|
64 |
+
# Calcular probabilidades
|
65 |
probabilities = torch.nn.functional.softmax(output[0], dim=0)
|
66 |
|
67 |
+
# Obter as 3 melhores previsões
|
68 |
top3_probs, top3_indices = torch.topk(probabilities, 3)
|
69 |
|
70 |
results = {
|
|
|
72 |
for p, i in zip(top3_probs, top3_indices)
|
73 |
}
|
74 |
|
75 |
+
# Obter a melhor previsão
|
76 |
best_class = class_names[top3_indices[0]]
|
77 |
best_conf = top3_probs[0].item() * 100
|
78 |
|
79 |
+
# Salvar resultados
|
80 |
with open('/tmp/prediction_results.txt', 'a') as f:
|
81 |
+
f.write(f"Imagem: {img}\n"
|
82 |
+
f"Previsão: {best_class}\n"
|
83 |
+
f"Confiança: {best_conf:.2f}%\n"
|
84 |
f"Top 3: {results}\n"
|
85 |
f"------------------------\n")
|
86 |
|
87 |
+
return best_class, f"{best_conf:.2f}%", results
|
88 |
|
89 |
+
# Criar interface Gradio
|
90 |
def create_interface():
|
91 |
examples = [
|
92 |
"r0_0_100.jpg",
|
93 |
"r0_18_100.jpg"
|
94 |
]
|
95 |
|
96 |
+
with gr.Blocks(title="Sistema de Classificação de Frutas", theme=gr.themes.Soft()) as demo:
|
97 |
+
gr.Markdown("# 🍎 Sistema de Reconhecimento de Frutas")
|
98 |
|
99 |
with gr.Row():
|
100 |
with gr.Column():
|
101 |
+
image_input = gr.Image(type="pil", label="Envie uma imagem")
|
102 |
gr.Examples(examples=examples, inputs=image_input)
|
103 |
+
submit_btn = gr.Button("Classificar", variant="primary")
|
104 |
|
105 |
with gr.Column():
|
106 |
+
best_pred = gr.Textbox(label="Resultado da Previsão")
|
107 |
+
confidence = gr.Textbox(label="Nível de Confiança")
|
108 |
full_results = gr.Label(label="Top 3", num_top_classes=3)
|
109 |
|
110 |
+
# Evento de clique do botão 'Classificar'
|
111 |
submit_btn.click(
|
112 |
fn=predict_image,
|
113 |
inputs=image_input,
|
|
|
119 |
|
120 |
if __name__ == "__main__":
|
121 |
interface = create_interface()
|
122 |
+
interface.launch(share=False)
|