Neural network implementation for gpu clusterization

[ChSonnabend]

This PR brings the neural network application to GPU clusterization in the reconstruction code. Several remarks inline:

• The code includes the ort_interface and GPUORTFloat16 classes in order to make it build, but they are also added in a separate PR
• The publishing of MC labels around each digit maximum should be changed still. Currently more labels are pushed than necessary, but it makes the code work and gives sensible output in the GPUQA output

[github-actions]

REQUEST FOR PRODUCTION RELEASES:
To request your PR to be included in production software, please add the corresponding labels called "async-" to your PR. Add the labels directly (if you have the permissions) or add a comment of the form (note that labels are separated by a ",")

+async-label <label1>, <label2>, !<label3> ...

This will add <label1> and <label2> and removes <label3>.

The following labels are available
async-2023-pbpb-apass4
async-2023-pp-apass4
async-2024-pp-apass1
async-2022-pp-apass7
async-2024-pp-cpass0

[ChSonnabend]

Ping @davidrohr