PyTorch-Pretrained-ViT

PyTorch-Pretrained-ViT copied to clipboard

How to do "Fine-Tuning" or "Feature-Extraction" in the model B_16 (or even L_16)?

2

Hi there, do you know how I can I use one of the two techniques above to do image classification on "Stanford Dogs Dataset"? I've already tried the "B_16_imagenet1k" model...

dgrnd4

Extract the transformer intermediate layer

1

I want extract the transformer intermediate layer. I use follow code， but it does not work. nn.Sequential(*list(model.children())， how should i do？

leolv131

Visualizing attention map

4

Hi. Does anyone know how we can have access to attention maps?

ParnianA

RuntimeError: view size is not compatible with input tensor's size and stride

2

Hi, When I try and run this with my own dataset with 4 classes, I get this error.  RuntimeError: view size is not...

aaliixii

General changes and adding of support for more functionality

2

Added support for 'H-14' and L'16' ViT models. Added support for downloading the models directly from Google's cloud storage. Corrected the Jax to Pytorch weights transformation. Previous methodology would lead...

arkel23

Hi, Thanks for this implementation. I saw the parameters of nn.Linear() are set to no_gard() in models.py Line:139. ``` python @torch.no_grad() def init_weights(self): def _init(m): if isinstance(m, nn.Linear): nn.init.xavier_uniform_(m.weight) #...

TianshengSun

self.pos_embedding error training on different dataset

6

I am using a pre-trained VIT model and trained on some different task but I got an error model.py file ``` class PositionalEmbedding1D(nn.Module): """Adds (optionally learned) positional embeddings to the...

khawar-islam

i got ``` free(): invalid size Aborted (core dumped) ``` when i tried the example code for the classification . how can i solve it, please

mathshangw

When Run [B_16, B_32, L_32 ] models, an error occurs because of position embedding.

1

First solved with "ViT(**image_size=image_size**)"

sdh5349

@lukemelas thanks for sharing code base , i had few queries 1.can we train crossvit for multilabel classification problem , if so what is the procedure 2. i have a...

abhigoku10