-
### Description
There's a bit more to this but overall this is the gist of it.
`describe` pushes to the `group_id_stack`. If a describe can't be handled a code lens will not be created, which in t…
-
### System Info
I've been examining the `interpolate_pos_encoding` method in the ViTMAE implementation, and I've discovered what appears to be an error in the calculation of `num_patches`. This error…
-
# 🐛 Bug
## Description
When doing defaults list interpolation in an inner config (nested folder) using an interpolation key from an the outermost config (base level config) it cannot find the interp…
-
hi,leeyeehoo,thank u for the released codes.
In image.py, I don't understand the meaning of the code" target = cv2.resize(target,(target.shape[1]/8,target.shape[0]/8),interpolation = cv2.INTER_CUBI…
-
Hi,
Thanks for sharing the work, when I try to run the vitl example in an A100 gpu, I found the inference time settles down to around 120ms rather than 13ms as stated in the repo, is there a reason…
-
I tried converting pretrained PyTorch model to ONNX using native torch support via torch.onnx.export but it seems like it does not support bicubic interpolation used in ViT architecture.
I see that y…
-
@effigies - is there any nibabel based code that does the same thing that `mri_convert -c` does?
satra updated
5 months ago
-
I've noticed that when I set interpolation order of `reg_aladin` to nearest neighbour or trilinear, the output image which was initially non-negative, may contain negatives. To me, this implied that t…
-
Currently you display the preview images in their original resolution, however they render very slow if the original images are 5333x4000px for example
So instead, please do cache some small versio…
-
As [tutorial](https://cs230-stanford.github.io/project-code-examples.html) says,for getting down-sized pics we should run build_database.py.However,a small problem occurred because Linux and Windows u…