hls4ml-rnn-toptag

This repository contains the training and validation codes relevant for the hls4ml RNN top-tag model

Training and Testing data

The binary classifier training and testing data are stored here: https://cernbox.cern.ch/index.php/s/0CBn5SsUPb5KDnX

• x_train.npy : Training input features. This is a 3D array. For each event there are 20 particles and for each particle there six variables.

• y_train.npy : Contains 5 class labels for each event. Use the 5th column for training and testing purpose.

• x_test.npy : Testing features. Same format as the training data.

• y_test.npy : Same as the training data. Use the 5th column for testing.

Trained models

Trainied models are stored inside the models directory

Small GRU models

model_toptag_gru.h5

• Total parameters: 231
• 1 GRU layer (5 units) + 1 Dense (5 nodes) + 1 output node

Small LSTM models

Training and testing scrips

All the training and testing scripts are inside the scripts directory