Spaces:
Runtime error
Runtime error
shukurullo2004
commited on
Commit
•
b221e1d
1
Parent(s):
1edd590
Upload 9 files
Browse files- app.py +72 -0
- examples/5612.jpg +0 -0
- examples/5903.jpg +0 -0
- examples/6561.jpg +0 -0
- examples/6652.jpg +0 -0
- examples/6800.jpg +0 -0
- melanoma_model1.pth +3 -0
- model.py +19 -0
- requirements.txt +3 -0
app.py
ADDED
@@ -0,0 +1,72 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gr
|
2 |
+
import os
|
3 |
+
import torch
|
4 |
+
from PIL import Image
|
5 |
+
from timeit import default_timer as timer
|
6 |
+
|
7 |
+
from model import create_model
|
8 |
+
from typing import Tuple, Dict
|
9 |
+
|
10 |
+
class_names = ['Benign', 'Malignant']
|
11 |
+
|
12 |
+
model, transform = create_model()
|
13 |
+
|
14 |
+
# Load saved weights
|
15 |
+
model.load_state_dict(
|
16 |
+
torch.load(
|
17 |
+
f="melanoma_model1.pth",
|
18 |
+
map_location=torch.device("cpu"), # load to CPU
|
19 |
+
)
|
20 |
+
)
|
21 |
+
|
22 |
+
|
23 |
+
### 3. Predict function ###
|
24 |
+
|
25 |
+
# Create predict function
|
26 |
+
|
27 |
+
def predict(img) -> Tuple[Dict, float]:
|
28 |
+
"""Transforms and performs a prediction on img and returns prediction and time taken.
|
29 |
+
"""
|
30 |
+
# Start the timer
|
31 |
+
start_time = timer()
|
32 |
+
|
33 |
+
# Apply transformations to the image
|
34 |
+
img_tensor = transform(img).unsqueeze(0).to(next(model.parameters()).device)
|
35 |
+
|
36 |
+
# Put model into evaluation mode
|
37 |
+
model.eval()
|
38 |
+
|
39 |
+
# Pass the image through the model
|
40 |
+
with torch.no_grad():
|
41 |
+
y_logits = model(img_tensor).squeeze()
|
42 |
+
y_pred_probs = torch.sigmoid(y_logits)
|
43 |
+
|
44 |
+
# Round the prediction probabilities to get binary predictions
|
45 |
+
y_pred_binary = torch.round(y_pred_probs).item()
|
46 |
+
|
47 |
+
# Create a dictionary with the class label and the corresponding prediction probability
|
48 |
+
pred_label = class_names[int(y_pred_binary)]
|
49 |
+
|
50 |
+
# Calculate the prediction time
|
51 |
+
pred_time = round(timer() - start_time, 5)
|
52 |
+
|
53 |
+
# Return the prediction dictionary and prediction time
|
54 |
+
return {pred_label: float(y_pred_probs)}, pred_time
|
55 |
+
|
56 |
+
# Create title, description and article strings
|
57 |
+
title = "Melanoma Cancer Detection"
|
58 |
+
description = "An Vision Tranformer feature extractor computer vision model to classify images of MELANOMA CANCER.."
|
59 |
+
article = " model is built by Shukurullo Meliboev using Kaggle's Melanoma disease datasets."
|
60 |
+
example_list = [["examples/" + example] for example in os.listdir("examples")]
|
61 |
+
# Create the Gradio demo
|
62 |
+
demo = gr.Interface(fn=predict, # mapping function from input to output
|
63 |
+
inputs=gr.Image(type="pil"), # what are the inputs?
|
64 |
+
outputs=[gr.Label(num_top_classes=1, label="Predictions"), # what are the outputs?
|
65 |
+
gr.Number(label="Prediction time (s)")], # our fn has two outputs, therefore we have two outputs
|
66 |
+
examples=example_list,
|
67 |
+
title=title,
|
68 |
+
description=description,
|
69 |
+
article=article)
|
70 |
+
|
71 |
+
# Launch the demo!
|
72 |
+
demo.launch(False) # generate a publically shareable URL?
|
examples/5612.jpg
ADDED
examples/5903.jpg
ADDED
examples/6561.jpg
ADDED
examples/6652.jpg
ADDED
examples/6800.jpg
ADDED
melanoma_model1.pth
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:cefce437c7b45b48513536454fcf2be41049995597180bb818a4a6b2ed0ae8d4
|
3 |
+
size 343259298
|
model.py
ADDED
@@ -0,0 +1,19 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import torch
|
2 |
+
import torchvision
|
3 |
+
from torch import nn
|
4 |
+
def create_model():
|
5 |
+
torch.manual_seed(42)
|
6 |
+
torch.cuda.manual_seed(42)
|
7 |
+
|
8 |
+
weights = torchvision.models.ViT_B_16_Weights.DEFAULT
|
9 |
+
transform = weights.transforms()
|
10 |
+
model = torchvision.models.vit_b_16(weights=weights)
|
11 |
+
|
12 |
+
# 4. Freeze all layers in base model
|
13 |
+
for param in model.parameters():
|
14 |
+
param.requires_grad = False
|
15 |
+
|
16 |
+
model.heads = nn.Sequential(
|
17 |
+
nn.Linear(768,1)
|
18 |
+
)
|
19 |
+
return model, transform
|
requirements.txt
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
torch==2.2.1
|
2 |
+
torchvision==0.17.1
|
3 |
+
gradio==4.22.0
|