`MetaTensor`: collate; decollate; dataset; dataloader; out=; indexing and iterating across batches by rijobro · Pull Request #4137 · Project-MONAI/MONAI

rijobro · 2022-04-14T14:38:07Z

Add collation and decollation for MetaTensor. Also add support for out= used as a kwarg, e.g., torch.add(a,b,out=c).

Non-breaking, so going into dev.

Status

Ready

Types of changes

Non-breaking change (fix or new feature that would not break existing functionality).
New tests added to cover the changes.

Signed-off-by: Richard Brown <[email protected]>

Nic-Ma

Hi @rijobro ,

Thanks for the quick update.
I am wondering whether it's possible to give a warning if user doesn't use our list_data_collate? Otherwise, I feel it is easily ignored and lead to unknown errors..

Thanks.

monai/data/utils.py

Nic-Ma · 2022-04-19T10:11:17Z

/build

Signed-off-by: Richard Brown <[email protected]>

…llate

rijobro · 2022-04-20T13:24:22Z

@ericspod @wyli @Nic-Ma I upgraded the logic so that a MetaTensor knows if it is a batch of data or not. This is so that indexing and iterating with batches of data return a subset of the metadata, as expected:

dl = DataLoader(ds)
batch = next(iter(dl))
batch[0]  # should only return the 0th image and 0th metadata
batch[:, -1]  # should return all metadata
batch[..., -1]  # should return all metadata
next(iter(batch))  # should return 0th image and 0th metadata.

Signed-off-by: Richard Brown <[email protected]>

rijobro · 2022-04-20T14:36:53Z

@Nic-Ma could you run /build?

Nic-Ma · 2022-04-20T14:42:35Z

Hi @rijobro ,

Thanks for the quick update. I am wondering whether it's possible to give a warning if user doesn't use our list_data_collate? Otherwise, I feel it is easily ignored and lead to unknown errors..

Thanks.

Hi @rijobro ,

Thanks for your update.
I am wondering whether it's possible to give a warning if user doesn't use our list_data_collate?
If can't, let's merge this PR for now.

Thanks.

Nic-Ma · 2022-04-20T14:46:47Z

/build

rijobro · 2022-04-20T14:49:54Z

From the following code snippet, the first is good (uses our collation fn). The middle case can have a warning. But how would we do the last case? Or are you just hoping for a warning in the middle case?

Edit: lambda x: x would also be ok as a collation function, as the data will be a list, each element is kept independent.

import torch
from monai.data.meta_tensor import MetaTensor
from monai.data.dataset import Dataset
from monai.data.dataloader import DataLoader
from torch.utils.data import DataLoader as TorchDataLoader

data = [MetaTensor(torch.rand((1, 20, 20))) for _ in range(5)]
ds = Dataset(data)

# good
dl = DataLoader(ds)
next(iter(dl))

# can report warning
dl_no_collate = DataLoader(ds, collate_fn=lambda x: x)
next(iter(dl_no_collate))

# can't report warning
torch_dl = TorchDataLoader(ds)
next(iter(torch_dl))

Nic-Ma · 2022-04-20T14:53:10Z

Hi @rijobro ,

Yeah, I was asking about the last case, I understand it's hard to give a warning there. Just worried about user experience.
If you don't have any ideas, I am OK to merge this PR first.

Thanks.

rijobro · 2022-04-20T14:55:14Z

I don't have any better ideas at the moment, but will certainly reflect on it. The MetaTensor docstring contains this info:

When creating a batch with this class, use monai.data.DataLoader as opposed to torch.utils.data.DataLoader, as this will take care of collating the metadata properly.

Add padding to filter to ensure same size after anti-aliasing Use replicate padding insteadof zero padding to avoid artifacts for non-zero boundary Reuse GaussianSmooth 4073 Enhance DynUNet doc-strings (Project-MONAI#4102) * Fix doc strings error Signed-off-by: Yiheng Wang <[email protected]> * remove duplicate places Signed-off-by: Yiheng Wang <[email protected]> 4105 drops pt16 support (Project-MONAI#4106) * update sys req Signed-off-by: Wenqi Li <[email protected]> * temp test Signed-off-by: Wenqi Li <[email protected]> * update code for torch>=1.7 Signed-off-by: Wenqi Li <[email protected]> * temp tests Signed-off-by: Wenqi Li <[email protected]> * fixes tests Signed-off-by: Wenqi Li <[email protected]> * autofix Signed-off-by: Wenqi Li <[email protected]> * fixes import Signed-off-by: Wenqi Li <[email protected]> * clear cache Signed-off-by: Wenqi Li <[email protected]> * update based on comments Signed-off-by: Wenqi Li <[email protected]> * remove temp cmd Signed-off-by: Wenqi Li <[email protected]> Make `pixelshuffle` scriptable (Project-MONAI#4109) * Update the existing functionality to comply with the `torchscript.jit.script` function. Signed-off-by: Ramon Emiliani <[email protected]> meta tensor (Project-MONAI#4077) * meta tensor Signed-off-by: Richard Brown <[email protected]> 4084 Add kwargs for `Tensor.to()` in engines (Project-MONAI#4112) * [DLMED] add kwargs for to() API Signed-off-by: Nic Ma <[email protected]> * [MONAI] python code formatting Signed-off-by: monai-bot <[email protected]> * [DLMED] fix typo Signed-off-by: Nic Ma <[email protected]> * [DLMED] fix flake8 Signed-off-by: Nic Ma <[email protected]> * [DLMED] update according to comments Signed-off-by: Nic Ma <[email protected]> Co-authored-by: monai-bot <[email protected]> fixes pytorch version tests (Project-MONAI#4127) Signed-off-by: Wenqi Li <[email protected]> update meta tensor api (Project-MONAI#4131) * update meta tensor api Signed-off-by: Wenqi Li <[email protected]> * update based on comments Signed-off-by: Wenqi Li <[email protected]> runtests.sh isort (Project-MONAI#4134) Signed-off-by: Richard Brown <[email protected]> update citation (Project-MONAI#4133) Signed-off-by: Wenqi Li <[email protected]> `ToMetaTensor` and `FromMetaTensor` transforms (Project-MONAI#4115) to and from meta no skip if before pytorch 1.7 (Project-MONAI#4139) * no skip if before pytorch 1.7 Signed-off-by: Richard Brown <[email protected]> * fix Signed-off-by: Richard Brown <[email protected]> * fix Signed-off-by: Richard Brown <[email protected]> [DLMED] fix file name in meta (Project-MONAI#4145) Signed-off-by: Nic Ma <[email protected]> 4116 Add support for advanced args of AMP (Project-MONAI#4132) * [DLMED] fix typo in bundle scripts Signed-off-by: Nic Ma <[email protected]> * [DLMED] add support for AMP args Signed-off-by: Nic Ma <[email protected]> * [MONAI] python code formatting Signed-off-by: monai-bot <[email protected]> * [DLMED] fix flake8 Signed-off-by: Nic Ma <[email protected]> Co-authored-by: monai-bot <[email protected]> New wsireader (Project-MONAI#4147) `MetaTensor`: collate; decollate; dataset; dataloader; out=; indexing and iterating across batches (Project-MONAI#4137) `MetaTensor`: collate; decollate; dataset; dataloader; out=; indexing and iterating across batches (Project-MONAI#4137)

collate , decollate, dataset, dataloader, out=

2622402

Signed-off-by: Richard Brown <[email protected]>

rijobro requested review from Nic-Ma and wyli April 14, 2022 14:38

rijobro added 4 commits April 14, 2022 15:39

mypy

19e68c9

Signed-off-by: Richard Brown <[email protected]>

skip decollation for pytorch 1.7

d017918

Signed-off-by: Richard Brown <[email protected]>

fix

b36cd10

Signed-off-by: Richard Brown <[email protected]>

fix

a8f0373

Signed-off-by: Richard Brown <[email protected]>

Nic-Ma approved these changes Apr 15, 2022

View reviewed changes

monai/data/utils.py Outdated Show resolved Hide resolved

Merge branch 'dev' into MetaTensor_collate_decollate

c6ae90d

rijobro added 2 commits April 20, 2022 14:20

add batch index testing

12afd4a

Signed-off-by: Richard Brown <[email protected]>

Merge remote-tracking branch 'MONAI/dev' into MetaTensor_collate_deco…

f7a583e

…llate

rijobro added 5 commits April 20, 2022 14:30

fixes

fb9b10f

Signed-off-by: Richard Brown <[email protected]>

fix

4875784

Signed-off-by: Richard Brown <[email protected]>

fix

b307d46

Signed-off-by: Richard Brown <[email protected]>

fix

e40553a

Signed-off-by: Richard Brown <[email protected]>

fix

f2c2548

Signed-off-by: Richard Brown <[email protected]>

rijobro enabled auto-merge (squash) April 20, 2022 14:38

rijobro changed the title ~~MetaTensor: collate , decollate, dataset, dataloader, out=~~ MetaTensor: collate; decollate; dataset; dataloader; out=; indexing and iterating across batches Apr 20, 2022

Nic-Ma disabled auto-merge April 20, 2022 14:46

rijobro merged commit 6dfb6a8 into Project-MONAI:dev Apr 20, 2022

rijobro deleted the MetaTensor_collate_decollate branch April 20, 2022 16:12

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

`MetaTensor`: collate; decollate; dataset; dataloader; out=; indexing and iterating across batches#4137

`MetaTensor`: collate; decollate; dataset; dataloader; out=; indexing and iterating across batches#4137
rijobro merged 13 commits intoProject-MONAI:devfrom
rijobro:MetaTensor_collate_decollate

rijobro commented Apr 14, 2022

Uh oh!

Nic-Ma left a comment

Uh oh!

Uh oh!

Nic-Ma commented Apr 19, 2022

Uh oh!

rijobro commented Apr 20, 2022

Uh oh!

rijobro commented Apr 20, 2022

Uh oh!

Nic-Ma commented Apr 20, 2022 •

edited

Loading

Uh oh!

Nic-Ma commented Apr 20, 2022

Uh oh!

rijobro commented Apr 20, 2022 •

edited

Loading

Uh oh!

Nic-Ma commented Apr 20, 2022

Uh oh!

rijobro commented Apr 20, 2022

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Conversation

rijobro commented Apr 14, 2022

Status

Types of changes

Uh oh!

Nic-Ma left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Nic-Ma commented Apr 19, 2022

Uh oh!

rijobro commented Apr 20, 2022

Uh oh!

rijobro commented Apr 20, 2022

Uh oh!

Nic-Ma commented Apr 20, 2022 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Nic-Ma commented Apr 20, 2022

Uh oh!

rijobro commented Apr 20, 2022 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Nic-Ma commented Apr 20, 2022

Uh oh!

rijobro commented Apr 20, 2022

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Nic-Ma commented Apr 20, 2022 •

edited

Loading

rijobro commented Apr 20, 2022 •

edited

Loading