Comments (2)
(cc @ericspod @john-zielke-snkeos follow-up of the discussions in #6854, self.R
and RandomState
is currently preferred over np.random.*
, this ticket is to track the issue.
e.g. preferred usage:
Line 65 in 8aabdc9
from monai.
If we wanted to first change the usages of np.random.*
to use a RandomState
then migrate in the future to Generator
we could do the following:
- Change the usage of
np.random.*
in transforms to useself.R
. This will change some expected deterministic behaviour but would be more reproducible. - In other places in core change routines to accept a
RandomState
object as an argument and use that, defaulting tonp.random
if none is provided. This would cover most use cases with minimal addition. @wyli 's code snippet is probably the technique to use for this. - In
apps/*
andauto3dseg
ask contributors what they would want to do, mentioning we're trying to tighten up the implementation with an eye to future migration.
We do need to have a more considered discussion about what to do about migration at some point, either we move to Generator
or we do implement our own class to replace RandomState
usage.
from monai.
Related Issues (20)
- RandSpatialCropSamplesd HOT 2
- Missing documentation for `Invert`
- Cannot use non-`nearest` interpolation methods with `Invertd`
- ResourceWarning: Implicitly cleaning up TemporaryDirectory
- Add deep supervision to classification models
- Support more dtypes in `pad_nd`
- The resolution to be used in input images for inferencing in a trained nodule detector model HOT 1
- How to get image file names from DataLoader HOT 1
- RandWeightedCropd in batch
- Cannot run V-Net on medical decathlon data HOT 6
- Numpy 2.0 Behaviour Changes HOT 1
- Adding Tailored Model Implementations into Generative Model Application
- Version 1.3.2 Hotfix For Numpy HOT 2
- Attribute error in utils type_conversion code in latest monai version 1.3.1 HOT 2
- Load data repeatedly in class CSVIterableDataset from monai/data/iterable_dataset.py
- releasing v1.3.2 -- checklist
- Adding Tailored ControlNet Implementations into Generative Model Application
- 'load_module()' deprecated in Python 3.12
- Add 'ignore_empty' parameter to loss functions for handling empty ground truth
- Ruff check failed
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from monai.