HP_NUM_UNITS = hp.HParam('num_units', hp.Discrete([10, 30]))
HP_NUM_UNITS2 = hp.HParam('num_units2', hp.Discrete([5, 15]))
HP_OPTIMIZER = hp.HParam('optimizer', hp.Discrete(['adam', 'sgd']))
METRIC_ACCURACY = 'recall_m'
with tf.summary.create_file_writer('logs/hparam_tuning').as_default():
hp.hparams_config(
hparams=[HP_NUM_UNITS,HP_NUM_UNITS2, HP_OPTIMIZER],
metrics=[hp.Metric(METRIC_ACCURACY, display_name='Recall')],
)
def train_test_model(hparams):
model = tf.keras.models.Sequential([
tf.keras.layers.Dense(hparams[HP_NUM_UNITS],input_dim=27,activation=tf.nn.relu),
tf.keras.layers.Dense(hparams[HP_NUM_UNITS2], activation=tf.nn.relu),
tf.keras.layers.Dense(1, activation=tf.nn.sigmoid)
])
model.compile(
optimizer=hparams[HP_OPTIMIZER],
loss='binary_crossentropy',
metrics=recall_m)
model.fit(x_train, y_train, epochs=10)
_, recall = model.evaluate(x_test, y_test)
return recall
def run(run_dir, hparams):
with tf.summary.create_file_writer(run_dir).as_default():
hp.hparams(hparams) # record the values used in this trial
recall = train_test_model(hparams)
tf.summary.scalar(METRIC_ACCURACY, recall, step=1)
session_num = 0
for num_units in HP_NUM_UNITS.domain.values:
for num_units2 in HP_NUM_UNITS2.domain.values:
for optimizer in HP_OPTIMIZER.domain.values:
hparams = {
HP_NUM_UNITS: num_units,
HP_NUM_UNITS2: num_units2,
HP_OPTIMIZER: optimizer,
}
run_name = "run-%d" % session_num
print('--- Starting trial: %s' % run_name)
print({h.name: hparams[h] for h in hparams})
run('logs/hparam_tuning/' + run_name, hparams)
session_num += 1
would you please help me? Im going to compile my model using the recall metric. but in hparams tab, I cant see any other metrics but accuracy. here is my cod:
HP_NUM_UNITS = hp.HParam('num_units', hp.Discrete([10, 30]))
HP_NUM_UNITS2 = hp.HParam('num_units2', hp.Discrete([5, 15]))
HP_OPTIMIZER = hp.HParam('optimizer', hp.Discrete(['adam', 'sgd']))
METRIC_ACCURACY = 'recall_m'
with tf.summary.create_file_writer('logs/hparam_tuning').as_default():
hp.hparams_config(
hparams=[HP_NUM_UNITS,HP_NUM_UNITS2, HP_OPTIMIZER],
metrics=[hp.Metric(METRIC_ACCURACY, display_name='Recall')],
)
def train_test_model(hparams):
model = tf.keras.models.Sequential([
tf.keras.layers.Dense(hparams[HP_NUM_UNITS],input_dim=27,activation=tf.nn.relu),
tf.keras.layers.Dense(hparams[HP_NUM_UNITS2], activation=tf.nn.relu),
tf.keras.layers.Dense(1, activation=tf.nn.sigmoid)
])
model.compile(
optimizer=hparams[HP_OPTIMIZER],
loss='binary_crossentropy',
metrics=recall_m)
model.fit(x_train, y_train, epochs=10)
_, recall = model.evaluate(x_test, y_test)
return recall
def run(run_dir, hparams):
with tf.summary.create_file_writer(run_dir).as_default():
hp.hparams(hparams) # record the values used in this trial
recall = train_test_model(hparams)
tf.summary.scalar(METRIC_ACCURACY, recall, step=1)
session_num = 0
for num_units in HP_NUM_UNITS.domain.values:
for num_units2 in HP_NUM_UNITS2.domain.values:
for optimizer in HP_OPTIMIZER.domain.values:
hparams = {
HP_NUM_UNITS: num_units,
HP_NUM_UNITS2: num_units2,
HP_OPTIMIZER: optimizer,
}
run_name = "run-%d" % session_num
print('--- Starting trial: %s' % run_name)
print({h.name: hparams[h] for h in hparams})
run('logs/hparam_tuning/' + run_name, hparams)
session_num += 1
Here is what I see in tensorboard
