batch_size=batch_size, shuffling=False, cache_path=None) for lr in learning_rates: for opt in optimizer_list: #### IF SOMETHING BREAKS PICKUP WHERE WE LEFT OFF if model_count < 152: print(f'Skipping: {model_count}') model_count += 1 continue #### if lr == 'custom': lr = CustomSchedule() if opt == 'rmsprop': opt_m = tf.keras.optimizers.RMSprop(lr) elif opt == 'adam': opt_m = tf.keras.optimizers.Adam(lr) elif opt == 'adagrad': opt_m = tf.keras.optimizers.Adagrad(lr) elif opt == 'adamax': opt_m = tf.keras.optimizers.Adamax(lr) elif opt == 'sgd': opt_m = tf.keras.optimizers.SGD(lr)
# ### Build Train/ Validation => Model Ready Input tf_train_ds = create_dataset(tf_train_ds, batch_size=batch_size, shuffling=True, cache_path=None) tf_valid_ds = create_dataset(tf_valid_ds, batch_size=batch_size, shuffling=False, cache_path=None) # ### Custom Learning Rate Scheduler #Example plt.style.use('ggplot') schedule = CustomSchedule() plt.plot(schedule(tf.range(25000, dtype=tf.float32))) plt.xlabel("Steps") plt.ylabel("Learning rate") # ### Setup Callbacks for Tensorboard # + start_profile_batch = steps + 10 stop_profile_batch = start_profile_batch + 100 profile_range = f"{start_profile_batch},{stop_profile_batch}" log_path = log_dir + "/" + datetime.datetime.now().strftime( "%Y-%m-%d_%H:%M:%S") tensorboard_callback = tf.keras.callbacks.TensorBoard( log_dir=log_path,
# ### Build Train/ Validation => Model Ready Input tf_train_ds = create_dataset(tf_train_ds, batch_size=batch_size, shuffling=True, cache_path=None) tf_valid_ds = create_dataset(tf_valid_ds, batch_size=batch_size, shuffling=False, cache_path=None) # ### Custom Learning Rate Scheduler #Example plt.style.use('ggplot') schedule = CustomSchedule() plt.plot(schedule(tf.range(25000, dtype=tf.float32))) plt.xlabel("Steps") plt.ylabel("Learning rate") # ### Setup Callbacks for Tensorboard # + start_profile_batch = steps + 10 stop_profile_batch = start_profile_batch + 100 profile_range = f"{start_profile_batch},{stop_profile_batch}" log_path = log_dir + "/" + datetime.datetime.now().strftime( "%Y-%m-%d_%H:%M:%S") tensorboard_callback = tf.keras.callbacks.TensorBoard( log_dir=log_path,