Severin Magel

Implementation of CNN for IceCube

1

Implement the CNN from [this paper](https://arxiv.org/abs/2101.11589) **Is your feature request related to a problem? Please describe.** Needed for direct comparison between the CNN and other architectures all within GraphNet trained...

This issue came up in the GraphNet Slack: When training with mixed precision and the `BinaryClassificationTask` with the `BinaryCrossEntropyLoss` you will get the following error: ``` RuntimeError: torch.nn.functional.binary_cross_entropy and torch.nn.BCELoss...

Treatment of hull entering tracks that are not found in the MCTree

3

**Describe the bug** This bug happens when using the `I3Calorimetry` Extractor. This extractor counts all energies that are present within a hull around the detector with a specified padding value....

Improving dataselection with strings

3

**Is your feature request related to a problem? Please describe.** When you want to select events from a database that all have a non-Nan value for a truth variable. Ideally,...

The function: https://github.com/graphnet-team/graphnet/blob/401f28b450e995cc567a70e83bb813757e9a4d63/src/graphnet/models/easy_model.py#L323 Can be really slow when querying for additional attributes, which is handled here: https://github.com/graphnet-team/graphnet/blob/401f28b450e995cc567a70e83bb813757e9a4d63/src/graphnet/models/easy_model.py#L381-L397 In the prediction loop, we are already looping through the whole data. Can...

Add cnn model

1

This is the big PR for the goal of adding CNN support to GraphNeT, enabling direct comparisons (see #771). The CNN support consists of: - ImageDefinition to represent data as...