edgenext_base.usi_in1k

Clasificación de imágenes

from urllib.request import urlopen
from PIL import Image
import timm

img = Image.open(urlopen(
'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png'
))

model = timm.create_model('edgenext_base.usi_in1k', pretrained=True)
model = model.eval()

# obtener transformaciones específicas del modelo (normalización, redimensionamiento)
data_config = timm.data.resolve_model_data_config(model)
transforms = timm.data.create_transform(**data_config, is_training=False)

output = model(transforms(img).unsqueeze(0)) # hacer batch de 1 imagen

top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5)

Extracción de mapas de características

from urllib.request import urlopen
from PIL import Image
import timm

img = Image.open(urlopen(
'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png'
))

model = timm.create_model(
'edgenext_base.usi_in1k',
pretrained=True,
features_only=True,
)
model = model.eval()

# obtener transformaciones específicas del modelo (normalización, redimensionamiento)
data_config = timm.data.resolve_model_data_config(model)
transforms = timm.data.create_transform(**data_config, is_training=False)

output = model(transforms(img).unsqueeze(0)) # hacer batch de 1 imagen

for o in output:
# imprimir forma de cada mapa de características en la salida
# ej.:
#  torch.Size([1, 80, 64, 64])
#  torch.Size([1, 160, 32, 32])
#  torch.Size([1, 288, 16, 16])
#  torch.Size([1, 584, 8, 8])

print(o.shape)

Embeddings de imágenes

from urllib.request import urlopen
from PIL import Image
import timm

img = Image.open(urlopen(
'https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/beignets-task-guide.png'
))

model = timm.create_model(
'edgenext_base.usi_in1k',
pretrained=True,
num_classes=0, # eliminar clasificador nn.Linear
)
model = model.eval()

# obtener transformaciones específicas del modelo (normalización, redimensionamiento)
data_config = timm.data.resolve_model_data_config(model)
transforms = timm.data.create_transform(**data_config, is_training=False)

output = model(transforms(img).unsqueeze(0)) # la salida es un tensor de forma (batch_size, num_features)

# o equivalentemente (sin necesidad de establecer num_classes=0)

output = model.forward_features(transforms(img).unsqueeze(0))
# la salida no está agrupada, es un tensor de forma (1, 584, 8, 8)

output = model.forward_head(output, pre_logits=True)
# la salida es un tensor de forma (1, num_features)

Citar

@inproceedings{Maaz2022EdgeNeXt,
title={EdgeNeXt: Efficiently Amalgamated CNN-Transformer Architecture for Mobile Vision Applications},
author={Muhammad Maaz and Abdelrahman Shaker and Hisham Cholakkal and Salman Khan and Syed Waqas Zamir and Rao Muhammad Anwer and Fahad Shahbaz Khan},
booktitle={International Workshop on Computational Aspects of Deep Learning at 17th European Conference on Computer Vision (CADL2022)},
year={2022},
organization={Springer}
}

@misc{https://doi.org/10.48550/arxiv.2204.03475,
doi = {10.48550/ARXIV.2204.03475},  
url = {https://arxiv.org/abs/2204.03475},  
author = {Ridnik, Tal and Lawen, Hussam and Ben-Baruch, Emanuel and Noy, Asaf},  
keywords = {Computer Vision and Pattern Recognition (cs.CV), Machine Learning (cs.LG), FOS: Computer and information sciences, FOS: Computer and information sciences},  
title = {Solving ImageNet: a Unified Scheme for Training any Backbone to Top Results},  
publisher = {arXiv},  
year = {2022},  
}