dask / dask/dask-ml

Skorch + HyperbandSearchCV Example

Open
#664 13 comments 0 reactions 0 assignees View on GitHub
Dominant language
Python
Stars
951
Forks
262
PR merge metrics
No merged PRs in 30d

Description

Can you post a working example that uses Skorch and HyperbandSearchCV? I haven't been able to find an actual working example.

The biggest challenge I've faced so far is how to determine the batch size being fed to the model. It's unclear if that is the chunk size and beyond that, it's unclear how Skorch's batching interacts with that.

If I run `search.fit(X, y)` with numpy arrays, the chunk size is massive and the grid search is very slow. If I try to chunk X, y in dask arrays, I get the following error:
```ValueError: With n_samples=1, test_size=0.058823529411764705 and train_size=0.9411764705882353, the resulting train set will be empty. Adjust any of the aforementioned parameters.```

Contributor guide

Open the contributing guide

Assessment

This issue has not been assessed yet.

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.