Comments (4)
If you pass a datamodule, it will only select the predefined validation loader and validate on that, but maybe I would like to validate on the train set and the validation set, for example when taking a pre-trained model and checking performance without training. Might also be relevant if you try something like cross validation, where you split your train/val sets. In my case, I am trying conformal prediction, where you need to take a subset of the validation set to create a separate calibration set and use the the model with that, so you need to control "which" split dataloader to apply validation to.
from torchgeo.
I can understand why you would want to be able to use a dataset if a data module doesn't exist, but why would you want to use a dataset if a data module does exist?
from torchgeo.
In order to do trainer.validate(model, dataloaders=datamodule.val_dataloader())
but not having to implement my own normalization scheme as a collate fn for every dataloader from a datamodule I want to use. So for example say I train one model and want to validate it on a bunch of datasets, then I could pass multiple dataloaders from different datasets or datamodules to trainer.validate()
from torchgeo.
But why not use trainer.validate(model, datamodule=datamodule)
for all data modules?
from torchgeo.
Related Issues (20)
- Switch from SMP to TorchSeg HOT 1
- Add plot method to IntersectionDataset HOT 1
- v0.5.2 missing PRs HOT 2
- Use ruff
- Add Inference Example HOT 1
- Switch coverage providers? HOT 1
- Auto download fails for FireRisk HOT 11
- Anomaly with RandomGrayScale tests HOT 2
- Add YAML formatter HOT 16
- Change documentation theme HOT 1
- CDL: cannot redownload additional years HOT 20
- Overrideable resample property for IntersectionDataset
- UnionDataset of two IntersectionDataset fails HOT 2
- RandomBatchGeoSampler produces nan or nodata values HOT 6
- Check if bbox of intersection is valid HOT 4
- Git clone and pip install results in 'Successfully installed UNKNOWN-0.0.0' HOT 10
- class_weights cannot be passed via config file as a tensor is expected HOT 5
- README.md benchmark dataset code HOT 17
- AgriFieldNet missing filename glob HOT 3
- CLI script for test HOT 2
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 torchgeo.