Introduction of vit_mediumd_patch16_reg4_gap_384.sbb2_e200_in12k_ft_in1k
Model Details of vit_mediumd_patch16_reg4_gap_384.sbb2_e200_in12k_ft_in1k
Model card for vit_mediumd_patch16_reg4_gap_384.sbb2_e200_in12k_ft_in1k
A Vision Transformer (ViT) image classification model. This is a
timm
specific variation of the architecture with registers, global average pooling.
There are a number of models in the lower end of model scales that originate in
timm
:
variant
width
mlp width (mult)
heads
depth
timm orig
tiny
192
768 (4)
3
12
n
wee
256
1280 (5)
4
14
y
pwee
256
1280 (5)
4
16 (parallel)
y
small
384
1536 (4)
6
12
n
little
320
1792 (5.6)
5
14
y
medium
512
2048 (4)
8
12
y
mediumd
512
2048 (4)
8
20
y
betwixt
640
2560 (4)
10
12
y
base
768
3072 (4)
12
12
n
Pretrained on ImageNet-12k and fine-tuned on ImageNet-1k by Ross Wightman in
timm
using recipe template described below.
Recipe details:
Searching for better baselines. Influced by Swin/DeiT/DeiT-III but w/ increased weight decay, moderate (in12k) to high (in1k) augmentation. Layer-decay used for fine-tune. Some runs used BCE and/or NAdamW instead of AdamW.
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('vit_mediumd_patch16_reg4_gap_384.sbb2_e200_in12k_ft_in1k', pretrained=True)
model = model.eval()
# get model specific transforms (normalization, resize)
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)) # unsqueeze single image into batch of 1
top5_probabilities, top5_class_indices = torch.topk(output.softmax(dim=1) * 100, k=5)
Feature Map Extraction
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(
'vit_mediumd_patch16_reg4_gap_384.sbb2_e200_in12k_ft_in1k',
pretrained=True,
features_only=True,
)
model = model.eval()
# get model specific transforms (normalization, resize)
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)) # unsqueeze single image into batch of 1for o in output:
# print shape of each feature map in output# e.g.:# torch.Size([1, 512, 24, 24])# torch.Size([1, 512, 24, 24])# torch.Size([1, 512, 24, 24])print(o.shape)
Image Embeddings
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(
'vit_mediumd_patch16_reg4_gap_384.sbb2_e200_in12k_ft_in1k',
pretrained=True,
num_classes=0, # remove classifier nn.Linear
)
model = model.eval()
# get model specific transforms (normalization, resize)
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)) # output is (batch_size, num_features) shaped tensor# or equivalently (without needing to set num_classes=0)
output = model.forward_features(transforms(img).unsqueeze(0))
# output is unpooled, a (1, 580, 512) shaped tensor
output = model.forward_head(output, pre_logits=True)
# output is a (1, num_features) shaped tensor
@misc{rw2019timm,
author = {Ross Wightman},
title = {PyTorch Image Models},
year = {2019},
publisher = {GitHub},
journal = {GitHub repository},
doi = {10.5281/zenodo.4414861},
howpublished = {\url{https://github.com/huggingface/pytorch-image-models}}
}
@article{darcet2023vision,
title={Vision Transformers Need Registers},
author={Darcet, Timoth{'e}e and Oquab, Maxime and Mairal, Julien and Bojanowski, Piotr},
journal={arXiv preprint arXiv:2309.16588},
year={2023}
}
@article{dosovitskiy2020vit,
title={An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale},
author={Dosovitskiy, Alexey and Beyer, Lucas and Kolesnikov, Alexander and Weissenborn, Dirk and Zhai, Xiaohua and Unterthiner, Thomas and Dehghani, Mostafa and Minderer, Matthias and Heigold, Georg and Gelly, Sylvain and Uszkoreit, Jakob and Houlsby, Neil},
journal={ICLR},
year={2021}
}
Runs of timm vit_mediumd_patch16_reg4_gap_384.sbb2_e200_in12k_ft_in1k on huggingface.co
44.5K
Total runs
0
24-hour runs
359
3-day runs
879
7-day runs
43.8K
30-day runs
More Information About vit_mediumd_patch16_reg4_gap_384.sbb2_e200_in12k_ft_in1k huggingface.co Model
More vit_mediumd_patch16_reg4_gap_384.sbb2_e200_in12k_ft_in1k license Visit here:
vit_mediumd_patch16_reg4_gap_384.sbb2_e200_in12k_ft_in1k huggingface.co is an AI model on huggingface.co that provides vit_mediumd_patch16_reg4_gap_384.sbb2_e200_in12k_ft_in1k's model effect (), which can be used instantly with this timm vit_mediumd_patch16_reg4_gap_384.sbb2_e200_in12k_ft_in1k model. huggingface.co supports a free trial of the vit_mediumd_patch16_reg4_gap_384.sbb2_e200_in12k_ft_in1k model, and also provides paid use of the vit_mediumd_patch16_reg4_gap_384.sbb2_e200_in12k_ft_in1k. Support call vit_mediumd_patch16_reg4_gap_384.sbb2_e200_in12k_ft_in1k model through api, including Node.js, Python, http.
vit_mediumd_patch16_reg4_gap_384.sbb2_e200_in12k_ft_in1k huggingface.co is an online trial and call api platform, which integrates vit_mediumd_patch16_reg4_gap_384.sbb2_e200_in12k_ft_in1k's modeling effects, including api services, and provides a free online trial of vit_mediumd_patch16_reg4_gap_384.sbb2_e200_in12k_ft_in1k, you can try vit_mediumd_patch16_reg4_gap_384.sbb2_e200_in12k_ft_in1k online for free by clicking the link below.
timm vit_mediumd_patch16_reg4_gap_384.sbb2_e200_in12k_ft_in1k online free url in huggingface.co:
vit_mediumd_patch16_reg4_gap_384.sbb2_e200_in12k_ft_in1k is an open source model from GitHub that offers a free installation service, and any user can find vit_mediumd_patch16_reg4_gap_384.sbb2_e200_in12k_ft_in1k on GitHub to install. At the same time, huggingface.co provides the effect of vit_mediumd_patch16_reg4_gap_384.sbb2_e200_in12k_ft_in1k install, users can directly use vit_mediumd_patch16_reg4_gap_384.sbb2_e200_in12k_ft_in1k installed effect in huggingface.co for debugging and trial. It also supports api for free installation.
vit_mediumd_patch16_reg4_gap_384.sbb2_e200_in12k_ft_in1k install url in huggingface.co: