improve the compatibility with non-CUDA environments #124
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Jim137
suggested changes
May 8, 2024
| @@ -123,7 +123,10 @@ def __init__(self, in_dim=3, out_dim=2, num=5, k=3, noise_scale=0.1, scale_base= | |||
| if isinstance(scale_base, float): | |||
| self.scale_base = torch.nn.Parameter(torch.ones(size, device=device) * scale_base).requires_grad_(sb_trainable) # make scale trainable | |||
| else: | |||
| self.scale_base = torch.nn.Parameter(torch.FloatTensor(scale_base).to(device)).requires_grad_(sb_trainable) | |||
| if torch.cuda.is_available(): | |||
There was a problem hiding this comment.
This could potentially lead to errors for users who have CUDA device but intend to run on CPU.
|
I do agree with @Jim137. BTW, if you already set device in the calling function there's no need to do so. If you want it to be transparent, even if I don't agree with this approach, you should modify EVERY method that regards cuda vs cpu. |
|
Was gonna make a pull request about this and then saw this - 100% agree! |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Many thanks to the author for proposing this amazing KAN. I modified the KANLayer.py file for better compatibility in non-CUDA environments. BTW, it is my first pull request. I would appreciate it if you could accept this request. @KindXiaoming
Before:
line 126 in KANLayer.py
self.scale_base = torch.nn.Parameter(torch.FloatTensor(scale_base)).requires_grad_(sb_trainable)After:
if torch.cuda.is_available(): self.scale_base = torch.nn.Parameter(torch.FloatTensor(scale_base).cuda()).requires_grad_(sb_trainable) else: self.scale_base = torch.nn.Parameter(torch.FloatTensor(scale_base)).requires_grad_(sb_trainable)