mlr3keras

mlr3keras copied to clipboard

Hi folks, any idea why this could be happening? Error in py_call_impl(callable, call_args$unnamed, call_args$named) : ValueError: ('Input has undefined `axis` dimension. Input shape: ', TensorShape([None, 54])) The dataset is formed...

igorgomes

### Description Hi, I was confused with when using "classif.kerascnn" method the "Feature Types" is "-". And the error is not supporting numeric feature type when training the learner. The...

MinWLi

Application of mlr3keras from a user perspective. Work in progress. - [ ] `README.md` - Links to learners do not work. - [ ] `README.md` - `reticulate::conda_create()` is called with...

be-marc

I was following steps for installation but got error: ``` > keras::install_keras("conda", tensorflow="2.3.1", envname="mlr3keras") Collecting package metadata (current_repodata.json): ...working... done Solving environment: ...working... done ## Package Plan ## environment location:...

MislavSag

Flexibility wrt tf and keras versions

1

Support multiple versions of keras and tensorflow

pfistfl

Compare to mlr3proba param spaces and implementation

3

pfistfl

validation data split from mlr3

1

We might wanna be able to access validation split performance

pfistfl

skorch

3

Should really look at skorch.readthedocs.io. I think this has many things we might want!

pfistfl

- [ ] Task to tfrecord - [ ] tfrecords backend https://keras.rstudio.com/articles/eager_guide.html#requirements

pfistfl

Perhaps also should be able to set a hyperparameter "weights" that allows me to set the weights. This could then be used for checkpointing and transfer learning etc.

pfistfl