-
I am using the latest version of autosklearn and running it on a high performance cluster.
Here is my code:
```
automl = autosklearn.classification.AutoSklearnClassifier(ensemble_size=1,
…
-
really like this feature. shouldnt it be possible to next to create a community driven db of ranges to also apply ml on general shape of the datasets (summary stats / topological data analysis) to pre…
-
Wonderful package.
My goal is to use it in parallel on multiple machines where they share the same storage.
To verify that the results will be reused from the different runs, I am running a small e…
-
The metafeatures generated for a dataset change depending on the target column, but the target column is not currently one of the metafeatures and is not being stored in the static knowledgebase metaf…
-
Right now, hooks registered need to be read-only, and cannot be used to modify the grad_input / output and still get correct results in the whole graph.
This is because we make some read-only assumpt…
-
```
/home/feurerm/projects/2018_nips_working_paper/2018_nips_working_paper/metalearning/greedy_portfolio.py:683: RuntimeWarning: invalid value encountered in greater_equal
np.all((normalized_matri…
-
This is on Windows, and I have very little idea what I am doing and what is happening:
I created a copy of a public database server that has 133 databases, because when trying to introspect on the …
-
OpenML is well known in the community, but D3M has the potential to add more datasets in the future. Using D3M would also allow us to be more useful for the D3M project.
-
Recently, we are trying to improve the ray log experience to better support user debugging.
Imaging that a user use the print function in his remote functions to output some debugging information. …
-
Is there a chance to investigate whether the One-Shot Reptile optimizer can deliver results comparable to Bayesian / Evolution?
I observed that the Bayesian / Evolution combo usually requires data…