ENH: Add log every nsteps parameter (#645)

Add log_every_n_steps trainer parameter

hi-ml/src/health_ml/deep_learning_config.py

@@ -531,6 +531,9 @@ class TrainerParams(param.Parameterized):
     pl_accumulate_grad_batches: int = param.Integer(default=1,
                                                     doc="The number of batches over which gradients are accumulated, "
                                                     "before a parameter update is done.")
+    pl_log_every_n_steps: int = param.Integer(default=50,
+                                              doc="PyTorch Lightning trainer flag 'log_every_n_steps': How often to"
+                                              "log within steps. Default to 50.")
 
     @property
     def use_gpu(self) -> bool:

hi-ml/src/health_ml/model_trainer.py

@@ -177,6 +177,7 @@ def create_lightning_trainer(container: LightningContainer,
                       limit_test_batches=container.pl_limit_test_batches or 1.0,
                       fast_dev_run=container.pl_fast_dev_run,  # type: ignore
                       num_sanity_val_steps=container.pl_num_sanity_val_steps,
+                      log_every_n_steps=container.pl_log_every_n_steps,
                       # check_val_every_n_epoch=container.pl_check_val_every_n_epoch,
                       callbacks=callbacks,
                       logger=loggers,