SUMIN/HuggingFace_transformer
1
0
Fork 0
Code Issues Pull Requests Actions 7 Packages Projects Releases Wiki Activity

Add speed metrics to all example scripts + template (#9260)

Browse Source
This commit is contained in:
Sylvain Gugger
2020-12-22 14:02:26 -05:00
committed by GitHub
parent 5b5f7dd09c
commit ab17758874
9 changed files with 118 additions and 19 deletions
Download Patch File Download Diff File Expand all files Collapse all files

15
examples/language-modeling/run_clm.py
View File

@@ -341,9 +341,20 @@ def main():
if (model_args.model_name_or_path is not None and os.path.isdir(model_args.model_name_or_path))
else None
)
trainer.train(model_path=model_path)
train_result = trainer.train(model_path=model_path)
trainer.save_model() # Saves the tokenizer too for easy upload
output_train_file = os.path.join(training_args.output_dir, "train_results.txt")
if trainer.is_world_process_zero():
with open(output_train_file, "w") as writer:
logger.info("***** Train results *****")
for key, value in sorted(train_result.metrics.items()):
logger.info(f" {key} = {value}")
writer.write(f"{key} = {value}\n")
# Need to save the state, since Trainer.save_model saves only the tokenizer with the model
trainer.state.save_to_json(os.path.join(training_args.output_dir, "trainer_state.json"))
# Evaluation
results = {}
if training_args.do_eval:
@@ -358,7 +369,7 @@ def main():
if trainer.is_world_process_zero():
with open(output_eval_file, "w") as writer:
logger.info("***** Eval results *****")
for key, value in results.items():
for key, value in sorted(results.items()):
logger.info(f" {key} = {value}")
writer.write(f"{key} = {value}\n")

15
examples/language-modeling/run_mlm.py
View File

@@ -376,9 +376,20 @@ def main():
if (model_args.model_name_or_path is not None and os.path.isdir(model_args.model_name_or_path))
else None
)
trainer.train(model_path=model_path)
train_result = trainer.train(model_path=model_path)
trainer.save_model() # Saves the tokenizer too for easy upload
output_train_file = os.path.join(training_args.output_dir, "train_results.txt")
if trainer.is_world_process_zero():
with open(output_train_file, "w") as writer:
logger.info("***** Train results *****")
for key, value in sorted(train_result.metrics.items()):
logger.info(f" {key} = {value}")
writer.write(f"{key} = {value}\n")
# Need to save the state, since Trainer.save_model saves only the tokenizer with the model
trainer.state.save_to_json(os.path.join(training_args.output_dir, "trainer_state.json"))
# Evaluation
results = {}
if training_args.do_eval:
@@ -393,7 +404,7 @@ def main():
if trainer.is_world_process_zero():
with open(output_eval_file, "w") as writer:
logger.info("***** Eval results *****")
for key, value in results.items():
for key, value in sorted(results.items()):
logger.info(f" {key} = {value}")
writer.write(f"{key} = {value}\n")

15
examples/language-modeling/run_mlm_wwm.py
View File

@@ -334,9 +334,20 @@ def main():
if (model_args.model_name_or_path is not None and os.path.isdir(model_args.model_name_or_path))
else None
)
trainer.train(model_path=model_path)
train_result = trainer.train(model_path=model_path)
trainer.save_model() # Saves the tokenizer too for easy upload
output_train_file = os.path.join(training_args.output_dir, "train_results.txt")
if trainer.is_world_process_zero():
with open(output_train_file, "w") as writer:
logger.info("***** Train results *****")
for key, value in sorted(train_result.metrics.items()):
logger.info(f" {key} = {value}")
writer.write(f"{key} = {value}\n")
# Need to save the state, since Trainer.save_model saves only the tokenizer with the model
trainer.state.save_to_json(os.path.join(training_args.output_dir, "trainer_state.json"))
# Evaluation
results = {}
if training_args.do_eval:
@@ -351,7 +362,7 @@ def main():
if trainer.is_world_process_zero():
with open(output_eval_file, "w") as writer:
logger.info("***** Eval results *****")
for key, value in results.items():
for key, value in sorted(results.items()):
logger.info(f" {key} = {value}")
writer.write(f"{key} = {value}\n")

15
examples/language-modeling/run_plm.py
View File

@@ -363,9 +363,20 @@ def main():
if (model_args.model_name_or_path is not None and os.path.isdir(model_args.model_name_or_path))
else None
)
trainer.train(model_path=model_path)
train_result = trainer.train(model_path=model_path)
trainer.save_model() # Saves the tokenizer too for easy upload
output_train_file = os.path.join(training_args.output_dir, "train_results.txt")
if trainer.is_world_process_zero():
with open(output_train_file, "w") as writer:
logger.info("***** Train results *****")
for key, value in sorted(train_result.metrics.items()):
logger.info(f" {key} = {value}")
writer.write(f"{key} = {value}\n")
# Need to save the state, since Trainer.save_model saves only the tokenizer with the model
trainer.state.save_to_json(os.path.join(training_args.output_dir, "trainer_state.json"))
# Evaluation
results = {}
if training_args.do_eval:
@@ -380,7 +391,7 @@ def main():
if trainer.is_world_process_zero():
with open(output_eval_file, "w") as writer:
logger.info("***** Eval results *****")
for key, value in results.items():
for key, value in sorted(results.items()):
logger.info(f" {key} = {value}")
writer.write(f"{key} = {value}\n")

15
examples/multiple-choice/run_swag.py
View File

@@ -317,11 +317,22 @@ def main():
# Training
if training_args.do_train:
trainer.train(
train_result = trainer.train(
model_path=model_args.model_name_or_path if os.path.isdir(model_args.model_name_or_path) else None
)
trainer.save_model() # Saves the tokenizer too for easy upload
output_train_file = os.path.join(training_args.output_dir, "train_results.txt")
if trainer.is_world_process_zero():
with open(output_train_file, "w") as writer:
logger.info("***** Train results *****")
for key, value in sorted(train_result.metrics.items()):
logger.info(f" {key} = {value}")
writer.write(f"{key} = {value}\n")
# Need to save the state, since Trainer.save_model saves only the tokenizer with the model
trainer.state.save_to_json(os.path.join(training_args.output_dir, "trainer_state.json"))
# Evaluation
results = {}
if training_args.do_eval:
@@ -333,7 +344,7 @@ def main():
if trainer.is_world_process_zero():
with open(output_eval_file, "w") as writer:
logger.info("***** Eval results *****")
for key, value in results.items():
for key, value in sorted(results.items()):
logger.info(f" {key} = {value}")
writer.write(f"{key} = {value}\n")

15
examples/question-answering/run_qa.py
View File

@@ -438,11 +438,22 @@ def main():
# Training
if training_args.do_train:
trainer.train(
train_result = trainer.train(
model_path=model_args.model_name_or_path if os.path.isdir(model_args.model_name_or_path) else None
)
trainer.save_model() # Saves the tokenizer too for easy upload
output_train_file = os.path.join(training_args.output_dir, "train_results.txt")
if trainer.is_world_process_zero():
with open(output_train_file, "w") as writer:
logger.info("***** Train results *****")
for key, value in sorted(train_result.metrics.items()):
logger.info(f" {key} = {value}")
writer.write(f"{key} = {value}\n")
# Need to save the state, since Trainer.save_model saves only the tokenizer with the model
trainer.state.save_to_json(os.path.join(training_args.output_dir, "trainer_state.json"))
# Evaluation
results = {}
if training_args.do_eval:
@@ -453,7 +464,7 @@ def main():
if trainer.is_world_process_zero():
with open(output_eval_file, "w") as writer:
logger.info("***** Eval results *****")
for key, value in results.items():
for key, value in sorted(results.items()):
logger.info(f" {key} = {value}")
writer.write(f"{key} = {value}\n")

15
examples/question-answering/run_qa_beam_search.py
View File

@@ -481,11 +481,22 @@ def main():
# Training
if training_args.do_train:
trainer.train(
train_result = trainer.train(
model_path=model_args.model_name_or_path if os.path.isdir(model_args.model_name_or_path) else None
)
trainer.save_model() # Saves the tokenizer too for easy upload
output_train_file = os.path.join(training_args.output_dir, "train_results.txt")
if trainer.is_world_process_zero():
with open(output_train_file, "w") as writer:
logger.info("***** Train results *****")
for key, value in sorted(train_result.metrics.items()):
logger.info(f" {key} = {value}")
writer.write(f"{key} = {value}\n")
# Need to save the state, since Trainer.save_model saves only the tokenizer with the model
trainer.state.save_to_json(os.path.join(training_args.output_dir, "trainer_state.json"))
# Evaluation
results = {}
if training_args.do_eval:
@@ -496,7 +507,7 @@ def main():
if trainer.is_world_process_zero():
with open(output_eval_file, "w") as writer:
logger.info("***** Eval results *****")
for key, value in results.items():
for key, value in sorted(results.items()):
logger.info(f" {key} = {value}")
writer.write(f"{key} = {value}\n")

15
examples/token-classification/run_ner.py
View File

@@ -340,11 +340,22 @@ def main():
# Training
if training_args.do_train:
trainer.train(
train_result = trainer.train(
model_path=model_args.model_name_or_path if os.path.isdir(model_args.model_name_or_path) else None
)
trainer.save_model() # Saves the tokenizer too for easy upload
output_train_file = os.path.join(training_args.output_dir, "train_results.txt")
if trainer.is_world_process_zero():
with open(output_train_file, "w") as writer:
logger.info("***** Train results *****")
for key, value in sorted(train_result.metrics.items()):
logger.info(f" {key} = {value}")
writer.write(f"{key} = {value}\n")
# Need to save the state, since Trainer.save_model saves only the tokenizer with the model
trainer.state.save_to_json(os.path.join(training_args.output_dir, "trainer_state.json"))
# Evaluation
results = {}
if training_args.do_eval:
@@ -377,7 +388,7 @@ def main():
output_test_results_file = os.path.join(training_args.output_dir, "test_results.txt")
if trainer.is_world_process_zero():
with open(output_test_results_file, "w") as writer:
for key, value in metrics.items():
for key, value in sorted(metrics.items()):
logger.info(f" {key} = {value}")
writer.write(f"{key} = {value}\n")

17
templates/adding_a_new_example_script/{{cookiecutter.directory_name}}/run_{{cookiecutter.example_shortcut}}.py
View File

@@ -308,7 +308,7 @@ def main():
# Training
if training_args.do_train:
{%- if cookiecutter.can_train_from_scratch == "False" %}
trainer.train(
train_result = trainer.train(
model_path=model_args.model_name_or_path if os.path.isdir(model_args.model_name_or_path) else None
)
{%- elif cookiecutter.can_train_from_scratch == "True" %}
@@ -317,10 +317,21 @@ def main():
if (model_args.model_name_or_path is not None and os.path.isdir(model_args.model_name_or_path))
else None
)
trainer.train(model_path=model_path)
train_result = trainer.train(model_path=model_path)
{% endif %}
trainer.save_model() # Saves the tokenizer too for easy upload
output_train_file = os.path.join(training_args.output_dir, "train_results.txt")
if trainer.is_world_process_zero():
with open(output_train_file, "w") as writer:
logger.info("***** Train results *****")
for key, value in sorted(train_result.metrics.items()):
logger.info(f" {key} = {value}")
writer.write(f"{key} = {value}\n")
# Need to save the state, since Trainer.save_model saves only the tokenizer with the model
trainer.state.save_to_json(os.path.join(training_args.output_dir, "trainer_state.json"))
# Evaluation
results = {}
if training_args.do_eval:
@@ -332,7 +343,7 @@ def main():
if trainer.is_world_process_zero():
with open(output_eval_file, "w") as writer:
logger.info("***** Eval results *****")
for key, value in results.items():
for key, value in sorted(results.items()):
logger.info(f" {key} = {value}")
writer.write(f"{key} = {value}\n")