def benchmarkTrainingBlockGRUVsGRUCell(self):
print("Comparison GRUBlockCell vs GRUCell")
print("--------------------------------------------------------------")
print("Training speed GRUBlockCell vs GRUCell")
print(
"batch_size, cell_size, input_size, time_steps, GPU, "
"basic_time_training, block_time_training, performance_training[%]"
)
iters = 10
for config in benchmarking.dict_product({
"use_gpu": [True, False],
"batch_size": [1, 32, 128],
"cell_size": [128, 512],
"input_size": [128, 512],
"time_steps": [50]
}):
basic_time, block_time = training_gru_block_vs_gru_cell(
config["batch_size"], config["cell_size"],
config["input_size"], config["time_steps"], config["use_gpu"],
iters)
self.report_benchmark(
name="GRUCell_training_time_BS%i_CS%i_IS%i_TS%i_gpu_%s" %
(config["batch_size"], config["cell_size"],
config["input_size"], config["time_steps"],
config["use_gpu"]),
iters=iters,
wall_time=basic_time)
self.report_benchmark(
name="GRUBlockCell_training_time_BS%i_CS%i_IS%i_TS%i_gpu_%s" %
(config["batch_size"], config["cell_size"],
config["input_size"], config["time_steps"],
config["use_gpu"]),
iters=iters,
wall_time=block_time)
def benchmarkSingleBpropStepBlockGRUVsGRUCell(self):
print("--------------------------------------------------------------")
print("Single bprop step speed GRUBlockCell vs GRUCell")
print("batch_size, cell_size, input_size, GPU, basic_time, "
"block_time, performance_inference[%]")
iters = 10
for config in benchmarking.dict_product({
"use_gpu": [True, False],
"batch_size": [1, 32, 128],
"cell_size": [128, 512],
"input_size": [128, 512]
}):
basic_time, block_time = single_bprop_step_gru_block_vs_gru_cell(
config["batch_size"], config["cell_size"],
config["input_size"], config["use_gpu"], iters)
self.report_benchmark(
name="GRUCell_Bprop_single_step_time_BS%i_CS%i_IS%i_gpu_%s" %
(config["batch_size"], config["cell_size"],
config["input_size"], config["use_gpu"]),
iters=iters,
wall_time=basic_time)
self.report_benchmark(
name="GRUBlockCell_Bprop_single_step_time_BS%i_CS%i_IS%i_gpu_%s"
% (config["batch_size"], config["cell_size"],
config["input_size"], config["use_gpu"]),
iters=iters,
wall_time=block_time)
print("--------------------------------------------------------------")
def benchmarkLSTMBlockCellFpropWithDynamicRNN(self):
print("BlockLSTMCell forward propagation via dynamic_rnn().")
print("--------------------------------------------------------------")
print("LSTMBlockCell Seconds per inference.")
print("batch_size,cell_size,input_size,time_steps,use_gpu,wall_time")
iters = 10
for config in benchmarking.dict_product({
"batch_size": [1, 8, 13, 32, 67, 128],
"cell_size": [128, 250, 512, 650, 1024, 1350],
"time_steps": [40],
"use_gpu": [True, False],
"dtype": ["float32", "float16"],
}):
dtype = dtypes.float32 if config[
"dtype"] == "float32" else dtypes.float16
with ops.Graph().as_default():
with benchmarking.device(use_gpu=config["use_gpu"]):
inputs = variable_scope.get_variable(
"x",
dtype=dtype,
shape=[
config["time_steps"], config["batch_size"],
config["cell_size"]
])
cell = lstm_ops.LSTMBlockCell(config["cell_size"],
dtype=dtype)
outputs = rnn.dynamic_rnn(cell,
inputs,
time_major=True,
dtype=dtype)
init_op = variables.global_variables_initializer()
with session.Session() as sess:
sess.run(init_op)
wall_time = benchmarking.seconds_per_run(
outputs, sess, iters)
# Print to stdout. If the TEST_REPORT_FILE_PREFIX environment variable
# is set, this will produce a copy-paste-able CSV file.
print(",".join(
map(str, [
config["dtype"], config["batch_size"],
config["cell_size"], config["cell_size"],
config["time_steps"], config["use_gpu"], wall_time
])))
benchmark_name_template = "_".join([
"LSTMBlockCell_fprop", "DT_%(dtype)s", "BS%(batch_size)i",
"CS%(cell_size)i", "IS%(cell_size)i", "TS%(time_steps)i",
"gpu_%(use_gpu)s"
])
self.report_benchmark(name=benchmark_name_template % config,
iters=iters,
wall_time=wall_time,
extras=config)