Support more dtypes in pad_nd

To be brief, I just have no idea why torch.bool (and other integer dtypes except for torch.int32) is considered to be not supported by PyTorch (it seems to work for me with PyTorch 2.3). I will be glad to add a PR.

MONAI/monai/transforms/croppad/functional.py

Lines 99 to 104 in 4029c42

    
           if mode in {"constant", "reflect", "edge", "replicate", "wrap", "circular"} and img.dtype not in { 
        
               torch.int16, 
        
               torch.int64, 
        
               torch.bool, 
        
               torch.uint8, 
        
           }:

@function2-llx

Hi @function2-llx @KumoLiu ,

I’d like to explore this issue and work on adding support for additional dtypes in
pad_nd. I’ve gone through the contributing guidelines, and since this issue is
labeled as “Contribution wanted”, I’d like to take it up.

This will be my first time working with the MONAI repository. I’ll validate the
behavior with recent PyTorch versions and follow up with a PR if everything looks
good.

fix: support more dtypes in pad_nd function

ENH: support additional dtypes in pad_nd (Project-MONAI#8672)

Navigation Menu

Repository navigation

Support more dtypes in `pad_nd` #7842

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

	if mode in {"constant", "reflect", "edge", "replicate", "wrap", "circular"} and img.dtype not in {
	torch.int16,
	torch.int64,
	torch.bool,
	torch.uint8,
	}:

Support more dtypes in pad_nd #7842

Description

Activity

shubham-61969 commented on Dec 24, 2025

Metadata

Metadata

Assignees

Labels

Type

Projects

Milestone

Relationships

Development

Issue actions

Support more dtypes in `pad_nd` #7842