Training Command

Command examples used to train models

1

cd <path_to_SIT_FUSE>/src/sit_fuse/train/
# Can be run outside of the repo via command line or in a script as well

2

# Pretrain encoder
python3 pretrain_encoder.py -y <path_to_yaml>
# E.g. set <path_to_yaml> to ../config/model/emas_fire_dbn_multi_layer_pl.yaml 

3

# Finetune encoder and train deep clustering MLP head
python3 finetune_dc_mlp_head.py -y <path_to_yaml>
# Same YAML file as above

4

# Train heirarchichal deep clustering MLP heads
python3 train_heirarchichal_deep_cluster.py -y <path_to_yaml>
# Same YAML file as above

Use the same YAML file as in Training Dataset Preprocessing