Philip Weigel

Graph construction before training

1

I've found that the calculations for constructing graphs can be a cpu bottleneck in training (especially when not many cpus are available). It would be very useful if the graphs...

For cases where the final layer of a model already reduces the dimension to the number of outputs, you may not want the affine layer in the `LearnedTask`. Added a...

When a `perturbation_dict` is provided to the graph definition with a setting for `charge`, it can lead to negative charges since it is sampled from a normal distribution (if you...

**Describe the bug** There seems to be a weird edge case when processing single file "datasets" using more than one worker. I guess this is because https://github.com/graphnet-team/graphnet/blob/65ec11c8ee5f5b8448343a2254df9d6a9bc6187c/src/graphnet/data/dataconverter.py#L277-L293 is setting `n_workers...

I3InferenceModule doesn't recognize a list of length 1

1

**Describe the bug** Using `I3InferenceModule` with a single task in a list results in a lot of spam. It checks if it's a list but doesn't check the length. I...