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

Update TF text classification example (#11496)

Browse Source 
Big refactor, fixes and multi-GPU/TPU support
This commit is contained in:
Matt
2021-04-30 13:45:33 +01:00
committed by GitHub
parent 8b945ef03e
commit 20d6931e32
3 changed files with 199 additions and 185 deletions
Download Patch File Download Diff File Expand all files Collapse all files

14
examples/tensorflow/text-classification/README.md
View File

@@ -54,6 +54,20 @@ After training, the model will be saved to `--output_dir`. Once your model is tr
by calling the script without a `--train_file` or `--validation_file`; simply pass it the output_dir containing by calling the script without a `--train_file` or `--validation_file`; simply pass it the output_dir containing
the trained model and a `--test_file` and it will write its predictions to a text file for you. the trained model and a `--test_file` and it will write its predictions to a text file for you.
### Multi-GPU and TPU usage
By default, the script uses a `MirroredStrategy` and will use multiple GPUs effectively if they are available. TPUs
can also be used by passing the name of the TPU resource with the `--tpu` argument.
### Memory usage and data loading
One thing to note is that all data is loaded into memory in this script. Most text classification datasets are small
enough that this is not an issue, but if you have a very large dataset you will need to modify the script to handle
data streaming. This is particularly challenging for TPUs, given the stricter requirements and the sheer volume of data
required to keep them fed. A full explanation of all the possible pitfalls is a bit beyond this example script and
README, but for more information you can see the 'Input Datasets' section of
[this document](https://www.tensorflow.org/guide/tpu).
### Example command ### Example command
``` ```
python run_text_classification.py \ python run_text_classification.py \

363
examples/tensorflow/text-classification/run_text_classification.py
View File

@@ -18,10 +18,8 @@
import logging import logging
import os import os
import random
import sys import sys
from dataclasses import dataclass, field from dataclasses import dataclass, field
from math import ceil
from pathlib import Path from pathlib import Path
from typing import Optional from typing import Optional
@@ -34,7 +32,7 @@ from transformers import (
HfArgumentParser, HfArgumentParser,
PretrainedConfig, PretrainedConfig,
TFAutoModelForSequenceClassification, TFAutoModelForSequenceClassification,
TrainingArguments, TFTrainingArguments,
set_seed, set_seed,
) )
from transformers.file_utils import CONFIG_NAME, TF2_WEIGHTS_NAME from transformers.file_utils import CONFIG_NAME, TF2_WEIGHTS_NAME
@@ -48,65 +46,6 @@ logger = logging.getLogger(__name__)
# region Helper classes # region Helper classes
class DataSequence(tf.keras.utils.Sequence):
# We use a Sequence object to load the data. Although it's completely possible to load your data as Numpy/TF arrays
# and pass those straight to the Model, this constrains you in a couple of ways. Most notably, it requires all
# the data to be padded to the length of the longest input example, and it also requires the whole dataset to be
# loaded into memory. If these aren't major problems for you, you can skip the sequence object in your own code!
def __init__(self, dataset, non_label_column_names, batch_size, labels, shuffle=True):
super().__init__()
# Retain all of the columns not present in the original data - these are the ones added by the tokenizer
self.data = {
key: dataset[key]
for key in dataset.features.keys()
if key not in non_label_column_names and key != "label"
}
data_lengths = {len(array) for array in self.data.values()}
assert len(data_lengths) == 1, "Dataset arrays differ in length!"
self.data_length = data_lengths.pop()
self.num_batches = ceil(self.data_length / batch_size)
if labels:
self.labels = np.array(dataset["label"])
assert len(self.labels) == self.data_length, "Labels not the same length as input arrays!"
else:
self.labels = None
self.batch_size = batch_size
self.shuffle = shuffle
if self.shuffle:
# Shuffle the data order
self.permutation = np.random.permutation(self.data_length)
else:
self.permutation = None
def on_epoch_end(self):
# If we're shuffling, reshuffle the data order after each epoch
if self.shuffle:
self.permutation = np.random.permutation(self.data_length)
def __getitem__(self, item):
# Note that this yields a batch, not a single sample
batch_start = item * self.batch_size
batch_end = (item + 1) * self.batch_size
if self.shuffle:
data_indices = self.permutation[batch_start:batch_end]
else:
data_indices = np.arange(batch_start, batch_end)
# We want to pad the data as little as possible, so we only pad each batch
# to the maximum length within that batch. We do that by stacking the variable-
# length inputs into a ragged tensor and then densifying it.
batch_input = {
key: tf.ragged.constant([data[i] for i in data_indices]).to_tensor() for key, data in self.data.items()
}
if self.labels is None:
return batch_input
else:
batch_labels = self.labels[data_indices]
return batch_input, batch_labels
def __len__(self):
return self.num_batches
class SavePretrainedCallback(tf.keras.callbacks.Callback): class SavePretrainedCallback(tf.keras.callbacks.Callback):
# Hugging Face models have a save_pretrained() method that saves both the weights and the necessary # Hugging Face models have a save_pretrained() method that saves both the weights and the necessary
# metadata to allow them to be loaded as a pretrained model in future. This is a simple Keras callback # metadata to allow them to be loaded as a pretrained model in future. This is a simple Keras callback
@@ -119,8 +58,50 @@ class SavePretrainedCallback(tf.keras.callbacks.Callback):
self.model.save_pretrained(self.output_dir) self.model.save_pretrained(self.output_dir)
def convert_dataset_for_tensorflow(
dataset, non_label_column_names, batch_size, dataset_mode="variable_batch", shuffle=True, drop_remainder=True
):
"""Converts a Hugging Face dataset to a Tensorflow Dataset. The dataset_mode controls whether we pad all batches
to the maximum sequence length, or whether we only pad to the maximum length within that batch. The former
is most useful when training on TPU, as a new graph compilation is required for each sequence length.
"""
def densify_ragged_batch(features, label=None):
features = {
feature: ragged_tensor.to_tensor(shape=batch_shape[feature]) for feature, ragged_tensor in features.items()
}
if label is None:
return features
else:
return features, label
feature_keys = list(set(dataset.features.keys()) - set(non_label_column_names + ["label"]))
if dataset_mode == "variable_batch":
batch_shape = {key: None for key in feature_keys}
data = {key: tf.ragged.constant(dataset[key]) for key in feature_keys}
elif dataset_mode == "constant_batch":
data = {key: tf.ragged.constant(dataset[key]) for key in feature_keys}
batch_shape = {
key: tf.concat(([batch_size], ragged_tensor.bounding_shape()[1:]), axis=0)
for key, ragged_tensor in data.items()
}
else:
raise ValueError("Unknown dataset mode!")
if "label" in dataset.features:
labels = tf.convert_to_tensor(np.array(dataset["label"]))
tf_dataset = tf.data.Dataset.from_tensor_slices((data, labels))
else:
tf_dataset = tf.data.Dataset.from_tensor_slices(data)
if shuffle:
tf_dataset = tf_dataset.shuffle(buffer_size=len(dataset))
tf_dataset = tf_dataset.batch(batch_size=batch_size, drop_remainder=drop_remainder).map(densify_ragged_batch)
return tf_dataset
# endregion # endregion
# region Command-line arguments # region Command-line arguments
@dataclass @dataclass
class DataTrainingArguments: class DataTrainingArguments:
@@ -155,6 +136,7 @@ class DataTrainingArguments:
metadata={ metadata={
"help": "Whether to pad all samples to `max_seq_length`. " "help": "Whether to pad all samples to `max_seq_length`. "
"If False, will pad the samples dynamically when batching to the maximum length in the batch." "If False, will pad the samples dynamically when batching to the maximum length in the batch."
"Data will always be padded when using TPUs."
}, },
) )
max_train_samples: Optional[int] = field( max_train_samples: Optional[int] = field(
@@ -164,17 +146,17 @@ class DataTrainingArguments:
"value if set." "value if set."
}, },
) )
max_eval_samples: Optional[int] = field( max_val_samples: Optional[int] = field(
default=None, default=None,
metadata={ metadata={
"help": "For debugging purposes or quicker training, truncate the number of evaluation examples to this " "help": "For debugging purposes or quicker training, truncate the number of validation examples to this "
"value if set." "value if set."
}, },
) )
max_predict_samples: Optional[int] = field( max_test_samples: Optional[int] = field(
default=None, default=None,
metadata={ metadata={
"help": "For debugging purposes or quicker training, truncate the number of predict examples to this " "help": "For debugging purposes or quicker training, truncate the number of test examples to this "
"value if set." "value if set."
}, },
) )
@@ -223,6 +205,7 @@ class ModelArguments:
"with private models)." "with private models)."
}, },
) )
tpu: Optional[str] = field(default=None, metadata={"help": "Name of the TPU resource to use, if available"})
# endregion # endregion
@@ -234,7 +217,7 @@ def main():
# or by passing the --help flag to this script. # or by passing the --help flag to this script.
# We now keep distinct sets of args, for a cleaner separation of concerns. # We now keep distinct sets of args, for a cleaner separation of concerns.
parser = HfArgumentParser((ModelArguments, DataTrainingArguments, TrainingArguments)) parser = HfArgumentParser((ModelArguments, DataTrainingArguments, TFTrainingArguments))
if len(sys.argv) == 2 and sys.argv[1].endswith(".json"): if len(sys.argv) == 2 and sys.argv[1].endswith(".json"):
# If we pass only one argument to the script and it's the path to a json file, # If we pass only one argument to the script and it's the path to a json file,
# let's parse it to get our arguments. # let's parse it to get our arguments.
@@ -322,12 +305,7 @@ def main():
is_regression = None is_regression = None
# endregion # endregion
# region Load pretrained model and tokenizer # region Load model config and tokenizer
# Set seed before initializing model
set_seed(training_args.seed)
#
# In distributed training, the .from_pretrained methods guarantee that only one local process can concurrently
# download model & vocab.
if checkpoint is not None: if checkpoint is not None:
config_path = training_args.output_dir config_path = training_args.output_dir
elif model_args.config_name: elif model_args.config_name:
@@ -355,34 +333,6 @@ def main():
revision=model_args.model_revision, revision=model_args.model_revision,
use_auth_token=True if model_args.use_auth_token else None, use_auth_token=True if model_args.use_auth_token else None,
) )
if checkpoint is None:
model_path = model_args.model_name_or_path
else:
model_path = checkpoint
model = TFAutoModelForSequenceClassification.from_pretrained(
model_path,
config=config,
cache_dir=model_args.cache_dir,
revision=model_args.model_revision,
use_auth_token=True if model_args.use_auth_token else None,
)
# endregion
# region Optimizer, loss and compilation
optimizer = tf.keras.optimizers.Adam(
learning_rate=training_args.learning_rate,
beta_1=training_args.adam_beta1,
beta_2=training_args.adam_beta2,
epsilon=training_args.adam_epsilon,
clipnorm=training_args.max_grad_norm,
)
if is_regression:
loss = tf.keras.losses.MeanSquaredError()
metrics = []
else:
loss = tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True)
metrics = ["accuracy"]
model.compile(optimizer=optimizer, loss=loss, metrics=metrics)
# endregion # endregion
# region Dataset preprocessing # region Dataset preprocessing
@@ -399,13 +349,6 @@ def main():
else: else:
sentence1_key, sentence2_key = non_label_column_names[0], None sentence1_key, sentence2_key = non_label_column_names[0], None
# Padding strategy
if data_args.pad_to_max_length:
padding = "max_length"
else:
# We will pad later, dynamically at batch creation, to the max sequence length in each batch
padding = False
if data_args.max_seq_length > tokenizer.model_max_length: if data_args.max_seq_length > tokenizer.model_max_length:
logger.warning( logger.warning(
f"The max_seq_length passed ({data_args.max_seq_length}) is larger than the maximum length for the" f"The max_seq_length passed ({data_args.max_seq_length}) is larger than the maximum length for the"
@@ -415,8 +358,8 @@ def main():
# Ensure that our labels match the model's, if it has some pre-specified # Ensure that our labels match the model's, if it has some pre-specified
if "train" in datasets: if "train" in datasets:
if not is_regression and model.config.label2id != PretrainedConfig(num_labels=num_labels).label2id: if not is_regression and config.label2id != PretrainedConfig(num_labels=num_labels).label2id:
label_name_to_id = model.config.label2id label_name_to_id = config.label2id
if list(sorted(label_name_to_id.keys())) == list(sorted(label_list)): if list(sorted(label_name_to_id.keys())) == list(sorted(label_list)):
label_to_id = label_name_to_id # Use the model's labels label_to_id = label_name_to_id # Use the model's labels
else: else:
@@ -431,15 +374,15 @@ def main():
else: else:
label_to_id = None label_to_id = None
# Now we've established our label2id, let's overwrite the model config with it. # Now we've established our label2id, let's overwrite the model config with it.
model.config.label2id = label_to_id config.label2id = label_to_id
if model.config.label2id is not None: if config.label2id is not None:
model.config.id2label = {id: label for label, id in label_to_id.items()} config.id2label = {id: label for label, id in label_to_id.items()}
else: else:
model.config.id2label = None config.id2label = None
else: else:
label_to_id = model.config.label2id # Just load the data from the model label_to_id = config.label2id # Just load the data from the model
if "validation" in datasets and model.config.label2id is not None: if "validation" in datasets and config.label2id is not None:
validation_label_list = datasets["validation"].unique("label") validation_label_list = datasets["validation"].unique("label")
for val_label in validation_label_list: for val_label in validation_label_list:
assert val_label in label_to_id, f"Label {val_label} is in the validation set but not the training set!" assert val_label in label_to_id, f"Label {val_label} is in the validation set but not the training set!"
@@ -449,87 +392,141 @@ def main():
args = ( args = (
(examples[sentence1_key],) if sentence2_key is None else (examples[sentence1_key], examples[sentence2_key]) (examples[sentence1_key],) if sentence2_key is None else (examples[sentence1_key], examples[sentence2_key])
) )
result = tokenizer(*args, padding=padding, max_length=max_seq_length, truncation=True) result = tokenizer(*args, max_length=max_seq_length, truncation=True)
# Map labels to IDs # Map labels to IDs
if model.config.label2id is not None and "label" in examples: if config.label2id is not None and "label" in examples:
result["label"] = [(model.config.label2id[l] if l != -1 else -1) for l in examples["label"]] result["label"] = [(config.label2id[l] if l != -1 else -1) for l in examples["label"]]
return result return result
datasets = datasets.map(preprocess_function, batched=True, load_from_cache_file=not data_args.overwrite_cache) datasets = datasets.map(preprocess_function, batched=True, load_from_cache_file=not data_args.overwrite_cache)
if "train" in datasets:
train_dataset = datasets["train"]
if data_args.max_train_samples is not None:
train_dataset = train_dataset.select(range(data_args.max_train_samples))
# Log a few random samples from the training set so we can see that it's working as expected:
for index in random.sample(range(len(train_dataset)), 3):
logger.info(f"Sample {index} of the training set: {train_dataset[index]}.")
if "validation" in datasets:
eval_dataset = datasets["validation"]
if data_args.max_eval_samples is not None:
eval_dataset = eval_dataset.select(range(data_args.max_eval_samples))
if "test" in datasets:
predict_dataset = datasets["test"]
if data_args.max_predict_samples is not None:
predict_dataset = predict_dataset.select(range(data_args.max_predict_samples))
# endregion # endregion
# region Training with training_args.strategy.scope():
if "train" in datasets: # region Load pretrained model
training_dataset = DataSequence( # Set seed before initializing model
train_dataset, non_label_column_names, batch_size=training_args.per_device_train_batch_size, labels=True set_seed(training_args.seed)
) #
if "validation" in datasets: # In distributed training, the .from_pretrained methods guarantee that only one local process can concurrently
eval_dataset = DataSequence( # download model & vocab.
eval_dataset, non_label_column_names, batch_size=training_args.per_device_eval_batch_size, labels=True if checkpoint is None:
) model_path = model_args.model_name_or_path
else: else:
eval_dataset = None model_path = checkpoint
model = TFAutoModelForSequenceClassification.from_pretrained(
model_path,
config=config,
cache_dir=model_args.cache_dir,
revision=model_args.model_revision,
use_auth_token=True if model_args.use_auth_token else None,
)
# endregion
callbacks = [SavePretrainedCallback(output_dir=training_args.output_dir)] # region Optimizer, loss and compilation
model.fit( optimizer = tf.keras.optimizers.Adam(
training_dataset, learning_rate=training_args.learning_rate,
validation_data=eval_dataset, beta_1=training_args.adam_beta1,
epochs=int(training_args.num_train_epochs), beta_2=training_args.adam_beta2,
callbacks=callbacks, epsilon=training_args.adam_epsilon,
clipnorm=training_args.max_grad_norm,
) )
elif "validation" in datasets:
# If there's a validation dataset but no training set, just evaluate the metrics
eval_dataset = DataSequence(
eval_dataset, non_label_column_names, batch_size=training_args.per_device_eval_batch_size, labels=True
)
logger.info("Computing metrics on validation data...")
if is_regression: if is_regression:
loss = model.evaluate(eval_dataset) loss_fn = tf.keras.losses.MeanSquaredError()
logger.info(f"Loss: {loss:.5f}") metrics = []
else: else:
loss, accuracy = model.evaluate(eval_dataset) loss_fn = tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True)
logger.info(f"Loss: {loss:.5f}, Accuracy: {accuracy * 100:.4f}%") metrics = ["accuracy"]
# endregion model.compile(optimizer=optimizer, loss=loss_fn, metrics=metrics)
# endregion
# region Prediction # region Convert data to TF format
if "test" in datasets:
logger.info("Doing predictions on Predict dataset...")
predict_dataset = DataSequence( # Convert data to a tf.keras.utils.Sequence object for training if we're not using a TPU
predict_dataset, non_label_column_names, batch_size=training_args.per_device_eval_batch_size, labels=False # For TPU, convert to a tf.data.Dataset
) tf_data = dict()
predictions = model.predict(predict_dataset)["logits"] max_samples = {
predictions = np.squeeze(predictions) if is_regression else np.argmax(predictions, axis=1) "train": data_args.max_train_samples,
output_predict_file = os.path.join(training_args.output_dir, "predict_results.txt") "validation": data_args.max_val_samples,
with open(output_predict_file, "w") as writer: "test": data_args.max_test_samples,
writer.write("index\tprediction\n") }
for index, item in enumerate(predictions): for key in ("train", "validation", "test"):
if is_regression: if key not in datasets:
writer.write(f"{index}\t{item:3.3f}\n") tf_data[key] = None
else: continue
item = model.config.id2label[item] if key in ("train", "validation"):
writer.write(f"{index}\t{item}\n") assert "label" in datasets[key].features, f"Missing labels from {key} data!"
logger.info(f"Wrote predictions to {output_predict_file}!") if key == "train":
shuffle = True
batch_size = training_args.per_device_train_batch_size
drop_remainder = True # Saves us worrying about scaling gradients for the last batch
else:
shuffle = False
batch_size = training_args.per_device_eval_batch_size
drop_remainder = False
samples_limit = max_samples[key]
dataset = datasets[key]
if samples_limit is not None:
dataset = dataset.select(range(samples_limit))
if isinstance(training_args.strategy, tf.distribute.TPUStrategy) or data_args.pad_to_max_length:
logger.info("Padding all batches to max length because argument was set or we're on TPU.")
dataset_mode = "constant_batch"
else:
dataset_mode = "variable_batch"
data = convert_dataset_for_tensorflow(
dataset,
non_label_column_names,
batch_size=batch_size,
dataset_mode=dataset_mode,
drop_remainder=drop_remainder,
shuffle=shuffle,
)
tf_data[key] = data
# endregion
# region Training and validation
if tf_data["train"] is not None:
callbacks = [SavePretrainedCallback(output_dir=training_args.output_dir)]
model.fit(
tf_data["train"],
validation_data=tf_data["validation"],
epochs=int(training_args.num_train_epochs),
callbacks=callbacks,
)
elif tf_data["validation"] is not None:
# If there's a validation dataset but no training set, just evaluate the metrics
logger.info("Computing metrics on validation data...")
if is_regression:
loss = model.evaluate(tf_data["validation"])
logger.info(f"Loss: {loss:.5f}")
else:
loss, accuracy = model.evaluate(tf_data["validation"])
logger.info(f"Loss: {loss:.5f}, Accuracy: {accuracy * 100:.4f}%")
# endregion
# region Prediction
if tf_data["test"] is not None:
logger.info("Doing predictions on test dataset...")
predictions = model.predict(tf_data["test"])["logits"]
predicted_class = np.squeeze(predictions) if is_regression else np.argmax(predictions, axis=1)
output_test_file = os.path.join(training_args.output_dir, "test_results.txt")
with open(output_test_file, "w") as writer:
writer.write("index\tprediction\n")
for index, item in enumerate(predicted_class):
if is_regression:
writer.write(f"{index}\t{item:3.3f}\n")
else:
item = config.id2label[item]
writer.write(f"{index}\t{item}\n")
logger.info(f"Wrote predictions to {output_test_file}!")
# endregion
# region Prediction losses
# This section is outside the scope() because it's very quick to compute, but behaves badly inside it
if "label" in datasets["test"].features:
print("Computing prediction loss on test labels...")
labels = datasets["test"]["label"]
loss = float(loss_fn(labels, predictions).numpy())
print(f"Test loss: {loss:.4f}")
# endregion # endregion

7
src/transformers/training_args_tf.py
View File

@@ -212,7 +212,10 @@ class TFTrainingArguments(TrainingArguments):
else: else:
tpu = tf.distribute.cluster_resolver.TPUClusterResolver() tpu = tf.distribute.cluster_resolver.TPUClusterResolver()
except ValueError: except ValueError:
tpu = None if self.tpu_name:
raise RuntimeError(f"Couldn't connect to TPU {self.tpu_name}!")
else:
tpu = None
if tpu: if tpu:
# Set to bfloat16 in case of TPU # Set to bfloat16 in case of TPU
@@ -233,7 +236,7 @@ class TFTrainingArguments(TrainingArguments):
# If you only want to use a specific subset of GPUs use `CUDA_VISIBLE_DEVICES=0` # If you only want to use a specific subset of GPUs use `CUDA_VISIBLE_DEVICES=0`
strategy = tf.distribute.MirroredStrategy() strategy = tf.distribute.MirroredStrategy()
else: else:
raise ValueError("Cannot find the proper strategy please check your environment properties.") raise ValueError("Cannot find the proper strategy, please check your environment properties.")
return strategy return strategy