lucidrains / vit-pytorch

Implementation of Vision Transformer, a simple way to achieve SOTA in vision classification with only a single transformer encoder, in Pytorch

lucidrains/vit-pytorch Issues

Attention maps for PiT
Updated 4 days ago
Question about example notebook
Closed 5 days ago1
Accessing last layer hidden states or embeddings for models like CrossViT, RegionViT (Extractor doesn't seem to work)
Closed a month ago8
what are the changes i should use for the large release?
Closed 20 days ago
EfficientFormer Request!
Updated 21 days ago
How to get the feature map of the vit encoder
Updated a month ago3
Global average pool mean is taking cls token into account as well
Updated a month ago
How to calculate Params and Flops for Vit?
Updated a month ago
Although there is no bug in this line, '1 n d' should be '1 1 d' in my opinion, it would be confused.
Closed a month ago2
MaxViT's MbConv doesn't match article.
Closed a month ago1
Add another MLP head in vision transformer
Updated 2 months ago2
Attention maps rectangular input
Closed 2 months ago
suggest any jupyter notebook or python code for use this library
Updated 3 months ago1
my own datasets? train.py pre.py?
Updated 3 months ago1
How to use Multi-Head Attention in ViT
Updated 3 months ago1
Masked Auto Encoder, class token and linear probing
Updated 3 months ago
Can anyone show me the example of Distillation please
Updated 3 months ago1
why not combine key and query linar layer into one
Updated 3 months ago1
issue VIT + Loss function
Updated 4 months ago
How to train custom dataset
Updated 4 months ago1
Did you miss dropout?
Closed 4 months ago2
CCT and non-square images
Closed 5 months ago2
A new idea
Updated 5 months ago
Question about code of `vit_for_small_dataset.py`
Closed 5 months ago2
Could you release tutorials and examples as many as possible? And is the .zip file from the COCO dataset?
Updated 5 months ago
Has a big time gap between GPU and CPU for Rearrange op.
Closed 6 months ago
Loading MAE pre-trained model for image classification fine tuning.
Updated 6 months ago
I want to use pretrained VITMAE to embed patches without any masks which inturn I plan to feed into a decoder, Can someone point me in the correct direction ?
Updated 6 months ago
Any chance to have conda installation?
Closed 6 months ago3
[BUG] efficient ViT does not support rectangular images
Closed 6 months ago2
RuntimeError: stack expects a non-empty TensorList
Closed 6 months ago2
MAE using pretrained VIT
Updated 7 months ago2
MAE image recon
Updated 7 months ago8
NesT implementation doesn't match the author/timm implementation
Closed 7 months ago3
About different patch embedding in ViT
Closed 7 months ago1
does vit-pytorch has mvit?
Closed 7 months ago1
Vit MAE decoder positional embeddings
Closed 7 months ago2
Vit MAE reconstruction size mismatch
Updated 7 months ago2
Fine-tuning without loading position embedding from pre-trained model
Updated 7 months ago
where is the train.py? tks
Updated 7 months ago
pre-trained model
Closed 8 months ago1
Training in cifar100
Updated 8 months ago1
Hot to get the features and positional embeddig information
Updated 8 months ago4
mae output normal size images
Updated 8 months ago1
batch_index_select for `x`, the `dim` should be set as 2?
Closed 8 months ago2
Vision Transformers for Federated Learning use-cases
Updated 8 months ago
MAE decoder pos_emb
Updated 8 months ago3
Vit3D for Brain CTscans
Updated 8 months ago6
Would it better swap the `masked_tokens` and `unmasked_tokens`?
Closed 9 months ago1
PiT doesn't have a channel argument
Closed 9 months ago1

ezoic increase your site revenue