Pytorch use tensorboard for hyperparameter

I use pytorch 1.3.1(cuda92py36hb0ba70e_0 ) for some classification models.
I would like to record all the parameters using tensorboard in particular I want to change all the parameters and create a single plot where I can see all the loss curves.
Is it possible? Any suggestion? Any tutorial on this?

thanks so much for any help

writer.add_scalar(
                        f"{EXPERIMENT_DIR}/test_loss_epoch", loss_test_avg, epoch,
                    )
                    writer.flush()