def test_predict_dataframe_with_feature_columns():
predictor = TensorflowPredictor(model_definition=build_model,
model_weights=weights)
data = pd.DataFrame([[1, 2], [3, 4]], columns=["A", "B"])
predictions = predictor.predict(data, feature_columns=["A"])
assert len(predictions) == 2
assert predictions.to_numpy().flatten().tolist() == [1, 3]
def test_predict_dataframe(use_gpu):
predictor = TensorflowPredictor(model_definition=build_model_multi_input,
use_gpu=use_gpu)
data_batch = pd.DataFrame({"A": [0.0, 0.0, 0.0], "B": [1.0, 2.0, 3.0]})
predictions = predictor.predict(data_batch)
assert len(predictions) == 3
assert predictions.to_numpy().flatten().tolist() == [1.0, 2.0, 3.0]
def test_predict_array(use_gpu):
predictor = TensorflowPredictor(model_definition=build_model,
model_weights=weights,
use_gpu=use_gpu)
data_batch = np.asarray([1, 2, 3])
predictions = predictor.predict(data_batch)
assert len(predictions) == 3
assert predictions.flatten().tolist() == [2, 4, 6]
def test_predict(batch_type):
predictor = TensorflowPredictor(model_definition=build_model_multi_input)
raw_batch = pd.DataFrame({"A": [0.0, 0.0, 0.0], "B": [1.0, 2.0, 3.0]})
data_batch = convert_pandas_to_batch_type(raw_batch,
type=TYPE_TO_ENUM[batch_type])
raw_predictions = predictor.predict(data_batch)
predictions = convert_batch_type_to_pandas(raw_predictions)
assert len(predictions) == 3
assert predictions.to_numpy().flatten().tolist() == [1.0, 2.0, 3.0]
def test_predict_array_with_preprocessor():
preprocessor = DummyPreprocessor()
predictor = TensorflowPredictor(model_definition=build_model,
preprocessor=preprocessor,
model_weights=weights)
data_batch = np.array([[1], [2], [3]])
predictions = predictor.predict(data_batch)
assert len(predictions) == 3
assert predictions.to_numpy().flatten().tolist() == [2, 4, 6]
assert hasattr(predictor.preprocessor, "_batch_transformed")
def test_predict_multi_output(use_gpu):
predictor = TensorflowPredictor(model_definition=build_model_multi_output,
use_gpu=use_gpu)
data_batch = np.array([1, 2, 3])
predictions = predictor.predict(data_batch)
# Model outputs two tensors
assert len(predictions) == 2
for k, v in predictions.items():
# Each tensor is of size 3
assert len(v) == 3
assert v.flatten().tolist() == [1, 2, 3]
def test_predict_array_with_input_shape_unspecified():
def model_definition():
return tf.keras.models.Sequential(
tf.keras.layers.Lambda(lambda tensor: tensor))
predictor = TensorflowPredictor(model_definition=model_definition,
model_weights=[])
data_batch = np.array([[1], [2], [3]])
predictions = predictor.predict(data_batch)
assert len(predictions) == 3
assert predictions.to_numpy().flatten().tolist() == [1, 2, 3]
def test_init():
preprocessor = DummyPreprocessor()
predictor = TensorflowPredictor(model_definition=build_model,
preprocessor=preprocessor,
model_weights=weights)
checkpoint = {MODEL_KEY: weights, PREPROCESSOR_KEY: preprocessor}
checkpoint_predictor = TensorflowPredictor.from_checkpoint(
Checkpoint.from_dict(checkpoint), build_model)
assert checkpoint_predictor.model_definition == predictor.model_definition
assert checkpoint_predictor.model_weights == predictor.model_weights
assert checkpoint_predictor.preprocessor == predictor.preprocessor
def test_predict_array_with_preprocessor(use_gpu):
preprocessor = DummyPreprocessor()
predictor = TensorflowPredictor(
model_definition=build_model,
preprocessor=preprocessor,
model_weights=weights,
use_gpu=use_gpu,
)
data_batch = np.array([1, 2, 3])
predictions = predictor.predict(data_batch)
assert len(predictions) == 3
assert predictions.flatten().tolist() == [4, 8, 12]
def test_predict_array():
checkpoint = {MODEL_KEY: weights}
predictor = TensorflowPredictor.from_checkpoint(
Checkpoint.from_dict(checkpoint), build_model)
data_batch = np.array([[1], [2], [3]])
predictions = predictor.predict(data_batch)
assert len(predictions) == 3
assert predictions.to_numpy().flatten().tolist() == [1, 2, 3]
def test_keras_callback_e2e():
epochs = 3
config = {
"epochs": epochs,
}
trainer = TensorflowTrainer(
train_loop_per_worker=train_func,
train_loop_config=config,
scaling_config=ScalingConfig(num_workers=2),
datasets={TRAIN_DATASET_KEY: get_dataset()},
)
checkpoint = trainer.fit().checkpoint
checkpoint_dict = checkpoint.to_dict()
assert MODEL_KEY in checkpoint_dict
predictor = TensorflowPredictor.from_checkpoint(
checkpoint, model_definition=build_model)
items = np.random.uniform(0, 1, size=(10, 1))
predictor.predict(data=items)
def __init__(self):
self.pred = TensorflowPredictor.from_checkpoint(
result.checkpoint, build_model)
TensorflowCheckpoint,
TensorflowPredictor,
)
def build_model() -> tf.keras.Model:
model = tf.keras.Sequential([
tf.keras.layers.InputLayer(input_shape=(1, )),
tf.keras.layers.Dense(1),
])
return model
model = build_model()
checkpoint = TensorflowCheckpoint.from_model(model)
predictor = TensorflowPredictor.from_checkpoint(checkpoint,
model_definition=build_model)
data = np.array([1, 2, 3, 4])
predictions = predictor.predict(data)
print(predictions)
# [[-1.6930283]
# [-3.3860567]
# [-5.079085 ]
# [-6.7721133]]
# __use_predictor_end__
# __batch_prediction_start__
import pandas as pd
from ray.train.batch_predictor import BatchPredictor
batch_predictor = BatchPredictor(checkpoint,