Is Fine - tuning feasible on the DTITR model?

[ChenaoB]

Hi

In the training task after creating the model

dtitr_model = build_dtitr_model(FLAGS, FLAGS.prot_transformer_depth[0], FLAGS.smiles_transformer_depth[0], FLAGS.cross_block_depth[0] ............... ............... FLAGS.out_mlp_depth[0], FLAGS.out_mlp_hdim[0], optimizer_fun)

Is it possible to load the weights of a previously model using the load_weights function and then train a new model with these initial parameters (do fine-tuning)?

dtitr_model.load_weights('Path')

Where Path is the address of the previous model.

Regards!