논문1 : EfficientNet: Rethinking Model Scaling for Convolutional Neural Networks / Youtube Link
논문2 : EfficientDet: Scalable and Efficient Object Detection / Youtube Link
분류 : LightWeight
저자 : Mingxing Tan, Ruoming Pang, Quoc V. Le
읽는 배경 : Recognition Basic. Understand confusing and ambiguous things.
읽으면서 생각할 포인트 : 코드와 함께 최대한 완벽히 이해하기. 이해한 것 정확히 기록해두기.
느낀점 :
목차
1. EfficientNet from youtube (바로가기)
2. Code - lukemelas/EfficientNet-PyTorch (바로가기)
3. EfficientDet from youtube (바로가기)
4. Code - zylo117/Yet-Another-EfficientDet-Pytorch (바로가기)

1. EfficientNet from youtube

2. lukemelas/EfficientNet-PyTorch

Github-Link : EfficientNet-PyTorch
EfficientNet은 a family of image classification models 이다. Based on MnasNet in term of AutoML, Compound Scaling.

Simply, Model 불러와 Classification 수행하기

import json
from PIL import Image
import torch
from torchvision import transforms

from efficientnet_pytorch import EfficientNet
model = EfficientNet.from_pretrained('efficientnet-b0')

# 0. Preprocess image
tfms = transforms.Compose([transforms.Resize(224), transforms.ToTensor(),
    transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225]),])
img = tfms(Image.open('img.jpg')).unsqueeze(0)
print(img.shape) # torch.Size([1, 3, 224, 224])

# 0. Load ImageNet class names
labels_map = json.load(open('labels_map.txt'))  # 이미 EfficientNet-Pytorch/examples/simple/labels_map.txt 있다.
labels_map = [labels_map[str(i)] for i in range(1000)]

# 1. For Classify
model.eval()
with torch.no_grad():
    outputs = model(img)

# 2. For Feature Extractor
features = model.extract_features(img)
print(features.shape) # torch.Size([1, 1280, 7, 7])