landmark_recognition / model_execute.py
molokhovdmitry's picture
Add main.py, change model_execute.py
732a846
raw
history blame
2.71 kB
import pickle
import torch
import torchvision.transforms as transforms
from PIL import Image
import csv
def preprocess_images(images):
"""
Preprocess image for the model.
"""
preprocess = transforms.Compose([
transforms.Resize([70, 70]),
transforms.ToTensor(),
transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]),
])
images_tensor = [preprocess(image) for image in images]
image_batch = torch.stack(images_tensor)
return image_batch
def output_to_names(output):
"""
Converts model outputs to category names names.
"""
with open('cat.csv') as file:
reader = csv.reader(file)
cat_list = list(reader)[0]
names = []
for prediction in output:
probabilities = torch.nn.functional.softmax(prediction, dim=0)
index = probabilities.argmax()
names.append(cat_list[index])
return names
def check_photo(name, photo):
preprocess = transforms.Compose([
transforms.Resize(256),
transforms.CenterCrop(224),
transforms.ToTensor(),
transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]),
])
input_tensor = preprocess(photo)
input_batch = input_tensor.unsqueeze(0) # create a mini-batch as expected by the model
# move the input and model to GPU for speed if available
if torch.cuda.is_available():
input_batch = input_batch.to('cuda')
model.to('cuda')
with torch.no_grad():
output = model(input_batch)
# Tensor of shape 1000, with confidence scores over ImageNet's 1000 classes
print(name, output[0])
# The output has unnormalized scores. To get probabilities, you can run a softmax on it.
probabilities = torch.nn.functional.softmax(output[0], dim=0)
print(name, probabilities)
if __name__ == "__main__":
pkl_filename = "pickle_model.pkl"
with open(pkl_filename, 'rb') as file:
model = pickle.load(file)
model.eval()
# sample execution (requires torchvision)
gates_photo = Image.open("gates500.jpg")
musk_photo = Image.open("mask.jpg")
bezos_photo = Image.open("bezos500.jpg")
zuker_photo = Image.open("zuckerberg500.jpg")
jobs_photo = Image.open("jobs500.jpg")
test_photos_dict = {'gates':gates_photo, 'musk':musk_photo, 'bezos':bezos_photo,'zuker': zuker_photo,'jobs': jobs_photo}
for name in test_photos_dict:
check_photo(name, test_photos_dict[name])
tensor = torch.tensor([[-1.8637, -1.6411, -1.5038, -2.9645, -1.8477, 6.5004], [-1.6067, -1.6597, -1.0925, 5.1295, -1.6491, -1.4739], [-0.2427, -0.6140, -1.1936, -2.1147, 4.8429, -2.0129]])
print(output_to_names(tensor))