sail-sg / volo

VOLO: Vision Outlooker for Visual Recognition

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

RuntimeError: The size of tensor a (28) must match the size of tensor b (14) at non-singleton dimension 2

zhang-pan opened this issue · comments

When use the pre-trained model VOLO-D4-448, the error as flow:
Traceback (most recent call last):
File "F:/volo-main/main1_all_complete.py", line 416, in
main()
File "F:/volo-main/main1_all_complete.py", line 168, in main
train_loss,train_accuracy=train(train_loader,model, loss_f,optimizer,epoch,args)
File "F:/volo-main/main1_all_complete.py", line 239, in train
logits,aux,auxx =model(image)
File "D:\Python36\lib\site-packages\torch\nn\modules\module.py", line 889, in _call_impl
result = self.forward(*input, **kwargs)
File "F:\volo-main\models\volo.py", line 614, in forward
x = self.forward_tokens(x)
File "F:\volo-main\models\volo.py", line 579, in forward_tokens
x = x + self.pos_embed
RuntimeError: The size of tensor a (28) must match the size of tensor b (14) at non-singleton dimension 2

Hi, did you fix it? I meet same problem.

class VOLO(nn.Module):
def init(self, layers, img_size=640, in_chans=3, patch_size=8, stem_hidden_dim=64,
embed_dims=None, num_heads=None, downsamples=None, outlook_attention=None,
mlp_ratios=None, qkv_bias=False, qk_scale=None, drop_rate=0., attn_drop_rate=0.,
drop_path_rate=0., norm_layer=nn.LayerNorm, post_layers=None, return_mean=False,
return_dense=True, mix_token=True, pooling_scale=2, out_kernel=3,
out_stride=2, out_padding=1):

image_size大小需要修改成你自己定义的!