This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Architecture | Model_I | Model_II | Model_III | |
---|---|---|---|---|
Convolutional Vision Transformer | 91.54% | 99.41% | 99.04% | |
CrossFormer | 91.20% | 97.42% | 98.13% | |
LeViT | 91.64% | 97.12% | 97.97% | |
TwinsSVT | 91.08% | 97.44% | 98.48% | |
CCT | 89.71% | 69.68% | 99.48% | |
CrossViT | 84.20% | 91.33% | 81.29% | |
CaiT | 67.93% | 62.97% | 69.38% | |
T2TViT | 88.12% | 94.33% | 77.29% | |
PiT | 40.63% | 33.60% | 34.18% |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
TWINSSVT_CONFIG = { | |
"network_type": "TwinsSVT", | |
"pretrained": False, | |
"image_size": 224, | |
"batch_size": 64, | |
"num_epochs": 15, | |
"optimizer_config": { | |
"name": "AdamW", | |
"weight_decay": 0.01, | |
"lr": 0.001, |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Hyperparameter | Value | |
---|---|---|
Batch size | 64 | |
Epochs | 15 | |
Image channels | 1 | |
Optimizer | AdamW | |
Weight decay | 0.01 | |
Learning rate schedule | Cosine schedule with warmup | |
Initial learning rate | 0.01 |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
from typing import Optional | |
from torchvision import transforms | |
from PIL import Image | |
import albumentations as A | |
from albumentations.pytorch import ToTensorV2 | |
from torch.utils.data import DataLoader, Dataset | |
def get_transform_train( | |
upsample_size: int, final_size: int, channels: Optional[int] = 1 |