2020-05-28 05:22:16 +08:00
|
|
|
import os
|
|
|
|
import tempfile
|
|
|
|
import unittest
|
|
|
|
from pathlib import Path
|
|
|
|
|
2020-06-08 21:31:12 +08:00
|
|
|
from transformers import AutoConfig, is_torch_available
|
2020-05-28 05:22:16 +08:00
|
|
|
|
2020-06-22 18:06:56 +08:00
|
|
|
from .utils import require_torch, torch_device
|
2020-05-28 05:22:16 +08:00
|
|
|
|
|
|
|
|
|
|
|
if is_torch_available():
|
|
|
|
from transformers import (
|
|
|
|
PyTorchBenchmarkArguments,
|
|
|
|
PyTorchBenchmark,
|
|
|
|
)
|
|
|
|
|
|
|
|
|
|
|
|
@require_torch
|
|
|
|
class BenchmarkTest(unittest.TestCase):
|
|
|
|
def check_results_dict_not_empty(self, results):
|
|
|
|
for model_result in results.values():
|
|
|
|
for batch_size, sequence_length in zip(model_result["bs"], model_result["ss"]):
|
|
|
|
result = model_result["result"][batch_size][sequence_length]
|
|
|
|
self.assertIsNotNone(result)
|
|
|
|
|
|
|
|
def test_inference_no_configs(self):
|
|
|
|
MODEL_ID = "sshleifer/tiny-gpt2"
|
|
|
|
benchmark_args = PyTorchBenchmarkArguments(
|
2020-06-22 18:06:56 +08:00
|
|
|
models=[MODEL_ID],
|
|
|
|
training=False,
|
|
|
|
no_inference=False,
|
|
|
|
sequence_lengths=[8],
|
|
|
|
batch_sizes=[1],
|
|
|
|
no_multi_process=True,
|
2020-05-28 05:22:16 +08:00
|
|
|
)
|
|
|
|
benchmark = PyTorchBenchmark(benchmark_args)
|
|
|
|
results = benchmark.run()
|
|
|
|
self.check_results_dict_not_empty(results.time_inference_result)
|
|
|
|
self.check_results_dict_not_empty(results.memory_inference_result)
|
|
|
|
|
2020-06-10 05:12:43 +08:00
|
|
|
def test_inference_torchscript(self):
|
|
|
|
MODEL_ID = "sshleifer/tiny-gpt2"
|
|
|
|
benchmark_args = PyTorchBenchmarkArguments(
|
|
|
|
models=[MODEL_ID],
|
|
|
|
training=False,
|
|
|
|
no_inference=False,
|
|
|
|
torchscript=True,
|
|
|
|
sequence_lengths=[8],
|
|
|
|
batch_sizes=[1],
|
2020-06-22 18:06:56 +08:00
|
|
|
no_multi_process=True,
|
|
|
|
)
|
|
|
|
benchmark = PyTorchBenchmark(benchmark_args)
|
|
|
|
results = benchmark.run()
|
|
|
|
self.check_results_dict_not_empty(results.time_inference_result)
|
|
|
|
self.check_results_dict_not_empty(results.memory_inference_result)
|
|
|
|
|
|
|
|
@unittest.skipIf(torch_device == "cpu", "Cant do half precision")
|
|
|
|
def test_inference_fp16(self):
|
|
|
|
MODEL_ID = "sshleifer/tiny-gpt2"
|
|
|
|
benchmark_args = PyTorchBenchmarkArguments(
|
|
|
|
models=[MODEL_ID],
|
|
|
|
training=False,
|
|
|
|
no_inference=False,
|
|
|
|
fp16=True,
|
|
|
|
sequence_lengths=[8],
|
|
|
|
batch_sizes=[1],
|
|
|
|
no_multi_process=True,
|
2020-06-10 05:12:43 +08:00
|
|
|
)
|
|
|
|
benchmark = PyTorchBenchmark(benchmark_args)
|
|
|
|
results = benchmark.run()
|
|
|
|
self.check_results_dict_not_empty(results.time_inference_result)
|
|
|
|
self.check_results_dict_not_empty(results.memory_inference_result)
|
|
|
|
|
2020-05-28 05:22:16 +08:00
|
|
|
def test_train_no_configs(self):
|
|
|
|
MODEL_ID = "sshleifer/tiny-gpt2"
|
|
|
|
benchmark_args = PyTorchBenchmarkArguments(
|
2020-06-22 18:06:56 +08:00
|
|
|
models=[MODEL_ID],
|
|
|
|
training=True,
|
|
|
|
no_inference=True,
|
|
|
|
sequence_lengths=[8],
|
|
|
|
batch_sizes=[1],
|
|
|
|
no_multi_process=True,
|
|
|
|
)
|
|
|
|
benchmark = PyTorchBenchmark(benchmark_args)
|
|
|
|
results = benchmark.run()
|
|
|
|
self.check_results_dict_not_empty(results.time_train_result)
|
|
|
|
self.check_results_dict_not_empty(results.memory_train_result)
|
|
|
|
|
|
|
|
@unittest.skipIf(torch_device == "cpu", "Cant do half precision")
|
|
|
|
def test_train_no_configs_fp16(self):
|
|
|
|
MODEL_ID = "sshleifer/tiny-gpt2"
|
|
|
|
benchmark_args = PyTorchBenchmarkArguments(
|
|
|
|
models=[MODEL_ID],
|
|
|
|
training=True,
|
|
|
|
no_inference=True,
|
|
|
|
sequence_lengths=[8],
|
|
|
|
batch_sizes=[1],
|
|
|
|
fp16=True,
|
|
|
|
no_multi_process=True,
|
2020-05-28 05:22:16 +08:00
|
|
|
)
|
|
|
|
benchmark = PyTorchBenchmark(benchmark_args)
|
|
|
|
results = benchmark.run()
|
|
|
|
self.check_results_dict_not_empty(results.time_train_result)
|
|
|
|
self.check_results_dict_not_empty(results.memory_train_result)
|
|
|
|
|
|
|
|
def test_inference_with_configs(self):
|
|
|
|
MODEL_ID = "sshleifer/tiny-gpt2"
|
2020-06-08 21:31:12 +08:00
|
|
|
config = AutoConfig.from_pretrained(MODEL_ID)
|
|
|
|
benchmark_args = PyTorchBenchmarkArguments(
|
2020-06-22 18:06:56 +08:00
|
|
|
models=[MODEL_ID],
|
|
|
|
training=False,
|
|
|
|
no_inference=False,
|
|
|
|
sequence_lengths=[8],
|
|
|
|
batch_sizes=[1],
|
|
|
|
no_multi_process=True,
|
2020-06-08 21:31:12 +08:00
|
|
|
)
|
|
|
|
benchmark = PyTorchBenchmark(benchmark_args, configs=[config])
|
|
|
|
results = benchmark.run()
|
|
|
|
self.check_results_dict_not_empty(results.time_inference_result)
|
|
|
|
self.check_results_dict_not_empty(results.memory_inference_result)
|
|
|
|
|
|
|
|
def test_inference_encoder_decoder_with_configs(self):
|
|
|
|
MODEL_ID = "sshleifer/tinier_bart"
|
|
|
|
config = AutoConfig.from_pretrained(MODEL_ID)
|
2020-05-28 05:22:16 +08:00
|
|
|
benchmark_args = PyTorchBenchmarkArguments(
|
2020-06-22 18:06:56 +08:00
|
|
|
models=[MODEL_ID],
|
|
|
|
training=False,
|
|
|
|
no_inference=False,
|
|
|
|
sequence_lengths=[8],
|
|
|
|
batch_sizes=[1],
|
|
|
|
no_multi_process=True,
|
2020-05-28 05:22:16 +08:00
|
|
|
)
|
|
|
|
benchmark = PyTorchBenchmark(benchmark_args, configs=[config])
|
|
|
|
results = benchmark.run()
|
|
|
|
self.check_results_dict_not_empty(results.time_inference_result)
|
|
|
|
self.check_results_dict_not_empty(results.memory_inference_result)
|
|
|
|
|
|
|
|
def test_train_with_configs(self):
|
2020-06-10 05:12:43 +08:00
|
|
|
MODEL_ID = "sshleifer/tiny-gpt2"
|
|
|
|
config = AutoConfig.from_pretrained(MODEL_ID)
|
|
|
|
benchmark_args = PyTorchBenchmarkArguments(
|
|
|
|
models=[MODEL_ID],
|
|
|
|
training=True,
|
|
|
|
no_inference=True,
|
|
|
|
sequence_lengths=[8],
|
|
|
|
batch_sizes=[1],
|
2020-06-22 18:06:56 +08:00
|
|
|
no_multi_process=True,
|
2020-06-10 05:12:43 +08:00
|
|
|
)
|
|
|
|
benchmark = PyTorchBenchmark(benchmark_args, configs=[config])
|
|
|
|
results = benchmark.run()
|
|
|
|
self.check_results_dict_not_empty(results.time_train_result)
|
2020-06-08 21:31:12 +08:00
|
|
|
self.check_results_dict_not_empty(results.memory_train_result)
|
|
|
|
|
|
|
|
def test_train_encoder_decoder_with_configs(self):
|
|
|
|
MODEL_ID = "sshleifer/tinier_bart"
|
|
|
|
config = AutoConfig.from_pretrained(MODEL_ID)
|
2020-05-28 05:22:16 +08:00
|
|
|
benchmark_args = PyTorchBenchmarkArguments(
|
2020-06-22 18:06:56 +08:00
|
|
|
models=[MODEL_ID],
|
|
|
|
training=True,
|
|
|
|
no_inference=True,
|
|
|
|
sequence_lengths=[8],
|
|
|
|
batch_sizes=[1],
|
|
|
|
no_multi_process=True,
|
2020-05-28 05:22:16 +08:00
|
|
|
)
|
|
|
|
benchmark = PyTorchBenchmark(benchmark_args, configs=[config])
|
|
|
|
results = benchmark.run()
|
|
|
|
self.check_results_dict_not_empty(results.time_train_result)
|
|
|
|
self.check_results_dict_not_empty(results.memory_train_result)
|
|
|
|
|
|
|
|
def test_save_csv_files(self):
|
|
|
|
MODEL_ID = "sshleifer/tiny-gpt2"
|
|
|
|
with tempfile.TemporaryDirectory() as tmp_dir:
|
|
|
|
benchmark_args = PyTorchBenchmarkArguments(
|
|
|
|
models=[MODEL_ID],
|
|
|
|
training=True,
|
|
|
|
no_inference=False,
|
|
|
|
save_to_csv=True,
|
|
|
|
sequence_lengths=[8],
|
|
|
|
batch_sizes=[1],
|
|
|
|
inference_time_csv_file=os.path.join(tmp_dir, "inf_time.csv"),
|
|
|
|
train_memory_csv_file=os.path.join(tmp_dir, "train_mem.csv"),
|
|
|
|
inference_memory_csv_file=os.path.join(tmp_dir, "inf_mem.csv"),
|
|
|
|
train_time_csv_file=os.path.join(tmp_dir, "train_time.csv"),
|
|
|
|
env_info_csv_file=os.path.join(tmp_dir, "env.csv"),
|
2020-06-22 18:06:56 +08:00
|
|
|
no_multi_process=True,
|
2020-05-28 05:22:16 +08:00
|
|
|
)
|
|
|
|
benchmark = PyTorchBenchmark(benchmark_args)
|
|
|
|
benchmark.run()
|
|
|
|
self.assertTrue(Path(os.path.join(tmp_dir, "inf_time.csv")).exists())
|
|
|
|
self.assertTrue(Path(os.path.join(tmp_dir, "train_time.csv")).exists())
|
|
|
|
self.assertTrue(Path(os.path.join(tmp_dir, "inf_mem.csv")).exists())
|
|
|
|
self.assertTrue(Path(os.path.join(tmp_dir, "train_mem.csv")).exists())
|
|
|
|
self.assertTrue(Path(os.path.join(tmp_dir, "env.csv")).exists())
|
2020-06-08 21:31:12 +08:00
|
|
|
|
|
|
|
def test_trace_memory(self):
|
|
|
|
MODEL_ID = "sshleifer/tiny-gpt2"
|
|
|
|
|
|
|
|
def _check_summary_is_not_empty(summary):
|
|
|
|
self.assertTrue(hasattr(summary, "sequential"))
|
|
|
|
self.assertTrue(hasattr(summary, "cumulative"))
|
|
|
|
self.assertTrue(hasattr(summary, "current"))
|
|
|
|
self.assertTrue(hasattr(summary, "total"))
|
|
|
|
|
|
|
|
with tempfile.TemporaryDirectory() as tmp_dir:
|
|
|
|
benchmark_args = PyTorchBenchmarkArguments(
|
|
|
|
models=[MODEL_ID],
|
|
|
|
training=True,
|
|
|
|
no_inference=False,
|
|
|
|
sequence_lengths=[8],
|
|
|
|
batch_sizes=[1],
|
|
|
|
log_filename=os.path.join(tmp_dir, "log.txt"),
|
|
|
|
log_print=True,
|
|
|
|
trace_memory_line_by_line=True,
|
2020-06-22 18:06:56 +08:00
|
|
|
no_multi_process=True,
|
2020-06-08 21:31:12 +08:00
|
|
|
)
|
|
|
|
benchmark = PyTorchBenchmark(benchmark_args)
|
|
|
|
result = benchmark.run()
|
|
|
|
_check_summary_is_not_empty(result.inference_summary)
|
|
|
|
_check_summary_is_not_empty(result.train_summary)
|
|
|
|
self.assertTrue(Path(os.path.join(tmp_dir, "log.txt")).exists())
|