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

Add LayoutLMv3 (#17060)

Browse Source 
* Make forward pass work

* More improvements

* Remove unused imports

* Remove timm dependency

* Improve loss calculation of token classifier

* Fix most tests

* Add docs

* Add model integration test

* Make all tests pass

* Add LayoutLMv3FeatureExtractor

* Improve integration test + make fixup

* Add example script

* Fix style

* Add LayoutLMv3Processor

* Fix style

* Add option to add visual labels

* Make more tokenizer tests pass

* Fix more tests

* Make more tests pass

* Fix bug and improve docs

* Fix import of processors

* Improve docstrings

* Fix toctree and improve docs

* Fix auto tokenizer

* Move tests to model folder

* Move tests to model folder

* change default behavior add_prefix_space

* add prefix space for fast

* add_prefix_spcae set to True for Fast

* no space before `unique_no_split` token

* add test to hightligh special treatment of added tokens

* fix `test_batch_encode_dynamic_overflowing` by building a long enough example

* fix `test_full_tokenizer` with add_prefix_token

* Fix tokenizer integration test

* Make the code more readable

* Add tests for LayoutLMv3Processor

* Fix style

* Add model to README and update init

* Apply suggestions from code review

* Replace asserts by value errors

* Add suggestion by @ducviet00

* Add model to doc tests

* Simplify script

* Improve README

* a step ahead to fix

* Update pair_input_test

* Make all tokenizer tests pass - phew

* Make style

* Add LayoutLMv3 to CI job

* Fix auto mapping

* Fix CI job name

* Make all processor tests pass

* Make tests of LayoutLMv2 and LayoutXLM consistent

* Add copied from statements to fast tokenizer

* Add copied from statements to slow tokenizer

* Remove add_visual_labels attribute

* Fix tests

* Add link to notebooks

* Improve docs of LayoutLMv3Processor

* Fix reference to section

Co-authored-by: SaulLu <lucilesaul.com@gmail.com>
Co-authored-by: Niels Rogge <nielsrogge@Nielss-MacBook-Pro.local>
This commit is contained in:
NielsRogge
2022-05-24 09:53:45 +02:00
committed by GitHub
parent 13541b4aa2
commit 31ee80d556
43 changed files with 8623 additions and 69 deletions
Download Patch File Download Diff File Expand all files Collapse all files

69
examples/research_projects/layoutlmv3/README.md Normal file
View File

@@ -0,0 +1,69 @@
<!---
Copyright 2022 The HuggingFace Team. All rights reserved.
Licensed under the Apache License, Version 2.0 (the "License");
you may not use this file except in compliance with the License.
You may obtain a copy of the License at
http://www.apache.org/licenses/LICENSE-2.0
Unless required by applicable law or agreed to in writing, software
distributed under the License is distributed on an "AS IS" BASIS,
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
See the License for the specific language governing permissions and
limitations under the License.
-->
# Token classification with LayoutLMv3 (PyTorch version)
This directory contains a script, `run_funsd_cord.py`, that can be used to fine-tune (or evaluate) LayoutLMv3 on form understanding datasets, such as [FUNSD](https://guillaumejaume.github.io/FUNSD/) and [CORD](https://github.com/clovaai/cord).
The script `run_funsd_cord.py` leverages the 🤗 Datasets library and the Trainer API. You can easily customize it to your needs.
## Fine-tuning on FUNSD
Fine-tuning LayoutLMv3 for token classification on [FUNSD](https://guillaumejaume.github.io/FUNSD/) can be done as follows:
```bash
python run_funsd_cord.py \
--model_name_or_path microsoft/layoutlmv3-base \
--dataset_name funsd \
--output_dir layoutlmv3-test \
--do_train \
--do_eval \
--max_steps 1000 \
--evaluation_strategy steps \
--eval_steps 100 \
--learning_rate 1e-5 \
--load_best_model_at_end \
--metric_for_best_model "eval_f1" \
--push_to_hub \
--push_to_hub°model_id layoutlmv3-finetuned-funsd
```
👀 The resulting model can be found here: https://huggingface.co/nielsr/layoutlmv3-finetuned-funsd. By specifying the `push_to_hub` flag, the model gets uploaded automatically to the hub (regularly), together with a model card, which includes metrics such as precision, recall and F1. Note that you can easily update the model card, as it's just a README file of the respective repo on the hub.
There's also the "Training metrics" [tab](https://huggingface.co/nielsr/layoutlmv3-finetuned-funsd/tensorboard), which shows Tensorboard logs over the course of training. Pretty neat, huh?
## Fine-tuning on CORD
Fine-tuning LayoutLMv3 for token classification on [CORD](https://github.com/clovaai/cord) can be done as follows:
```bash
python run_funsd_cord.py \
--model_name_or_path microsoft/layoutlmv3-base \
--dataset_name cord \
--output_dir layoutlmv3-test \
--do_train \
--do_eval \
--max_steps 1000 \
--evaluation_strategy steps \
--eval_steps 100 \
--learning_rate 5e-5 \
--load_best_model_at_end \
--metric_for_best_model "eval_f1" \
--push_to_hub \
--push_to_hub°model_id layoutlmv3-finetuned-cord
```
👀 The resulting model can be found here: https://huggingface.co/nielsr/layoutlmv3-finetuned-cord. Note that a model card gets generated automatically in case you specify the `push_to_hub` flag.

2
examples/research_projects/layoutlmv3/requirements.txt Normal file
View File

@@ -0,0 +1,2 @@
datasets
seqeval

533
examples/research_projects/layoutlmv3/run_funsd_cord.py Normal file
View File

@@ -0,0 +1,533 @@
#!/usr/bin/env python
# coding=utf-8
# Copyright 2022 The HuggingFace Team All rights reserved.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
"""
Fine-tuning LayoutLMv3 for token classification on FUNSD or CORD.
"""
# You can also adapt this script on your own token classification task and datasets. Pointers for this are left as
# comments.
import logging
import os
import sys
from dataclasses import dataclass, field
from typing import Optional
import datasets
import numpy as np
from datasets import ClassLabel, load_dataset, load_metric
import transformers
from transformers import (
AutoConfig,
AutoModelForTokenClassification,
AutoProcessor,
HfArgumentParser,
Trainer,
TrainingArguments,
set_seed,
)
from transformers.data.data_collator import default_data_collator
from transformers.trainer_utils import get_last_checkpoint
from transformers.utils import check_min_version
from transformers.utils.versions import require_version
# Will error if the minimal version of Transformers is not installed. Remove at your own risks.
check_min_version("4.19.0.dev0")
require_version("datasets>=1.8.0", "To fix: pip install -r examples/pytorch/token-classification/requirements.txt")
logger = logging.getLogger(__name__)
@dataclass
class ModelArguments:
"""
Arguments pertaining to which model/config/tokenizer we are going to fine-tune from.
"""
model_name_or_path: str = field(
default="microsoft/layoutlmv3-base",
metadata={"help": "Path to pretrained model or model identifier from huggingface.co/models"},
)
config_name: Optional[str] = field(
default=None, metadata={"help": "Pretrained config name or path if not the same as model_name"}
)
processor_name: Optional[str] = field(
default=None, metadata={"help": "Name or path to the processor files if not the same as model_name"}
)
cache_dir: Optional[str] = field(
default=None,
metadata={"help": "Where do you want to store the pretrained models downloaded from huggingface.co"},
)
model_revision: str = field(
default="main",
metadata={"help": "The specific model version to use (can be a branch name, tag name or commit id)."},
)
use_auth_token: bool = field(
default=False,
metadata={
"help": (
"Will use the token generated when running `transformers-cli login` (necessary to use this script "
"with private models)."
)
},
)
@dataclass
class DataTrainingArguments:
"""
Arguments pertaining to what data we are going to input our model for training and eval.
"""
task_name: Optional[str] = field(default="ner", metadata={"help": "The name of the task (ner, pos...)."})
dataset_name: Optional[str] = field(
default="nielsr/funsd-layoutlmv3",
metadata={"help": "The name of the dataset to use (via the datasets library)."},
)
dataset_config_name: Optional[str] = field(
default=None, metadata={"help": "The configuration name of the dataset to use (via the datasets library)."}
)
train_file: Optional[str] = field(
default=None, metadata={"help": "The input training data file (a csv or JSON file)."}
)
validation_file: Optional[str] = field(
default=None,
metadata={"help": "An optional input evaluation data file to evaluate on (a csv or JSON file)."},
)
test_file: Optional[str] = field(
default=None,
metadata={"help": "An optional input test data file to predict on (a csv or JSON file)."},
)
text_column_name: Optional[str] = field(
default=None, metadata={"help": "The column name of text to input in the file (a csv or JSON file)."}
)
label_column_name: Optional[str] = field(
default=None, metadata={"help": "The column name of label to input in the file (a csv or JSON file)."}
)
overwrite_cache: bool = field(
default=False, metadata={"help": "Overwrite the cached training and evaluation sets"}
)
preprocessing_num_workers: Optional[int] = field(
default=None,
metadata={"help": "The number of processes to use for the preprocessing."},
)
max_seq_length: int = field(
default=512,
metadata={
"help": (
"The maximum total input sequence length after tokenization. If set, sequences longer "
"than this will be truncated, sequences shorter will be padded."
)
},
)
max_train_samples: Optional[int] = field(
default=None,
metadata={
"help": (
"For debugging purposes or quicker training, truncate the number of training examples to this "
"value if set."
)
},
)
max_eval_samples: Optional[int] = field(
default=None,
metadata={
"help": (
"For debugging purposes or quicker training, truncate the number of evaluation examples to this "
"value if set."
)
},
)
max_predict_samples: Optional[int] = field(
default=None,
metadata={
"help": (
"For debugging purposes or quicker training, truncate the number of prediction examples to this "
"value if set."
)
},
)
label_all_tokens: bool = field(
default=False,
metadata={
"help": (
"Whether to put the label for one word on all tokens of generated by that word or just on the "
"one (in which case the other tokens will have a padding index)."
)
},
)
return_entity_level_metrics: bool = field(
default=False,
metadata={"help": "Whether to return all the entity levels during evaluation or just the overall ones."},
)
def __post_init__(self):
if self.dataset_name is None and self.train_file is None and self.validation_file is None:
raise ValueError("Need either a dataset name or a training/validation file.")
else:
if self.train_file is not None:
extension = self.train_file.split(".")[-1]
assert extension in ["csv", "json"], "`train_file` should be a csv or a json file."
if self.validation_file is not None:
extension = self.validation_file.split(".")[-1]
assert extension in ["csv", "json"], "`validation_file` should be a csv or a json file."
self.task_name = self.task_name.lower()
def main():
# See all possible arguments in src/transformers/training_args.py
# or by passing the --help flag to this script.
# We now keep distinct sets of args, for a cleaner separation of concerns.
parser = HfArgumentParser((ModelArguments, DataTrainingArguments, TrainingArguments))
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,
# let's parse it to get our arguments.
model_args, data_args, training_args = parser.parse_json_file(json_file=os.path.abspath(sys.argv[1]))
else:
model_args, data_args, training_args = parser.parse_args_into_dataclasses()
# Setup logging
logging.basicConfig(
format="%(asctime)s - %(levelname)s - %(name)s - %(message)s",
datefmt="%m/%d/%Y %H:%M:%S",
handlers=[logging.StreamHandler(sys.stdout)],
)
log_level = training_args.get_process_log_level()
logger.setLevel(log_level)
datasets.utils.logging.set_verbosity(log_level)
transformers.utils.logging.set_verbosity(log_level)
transformers.utils.logging.enable_default_handler()
transformers.utils.logging.enable_explicit_format()
# Log on each process the small summary:
logger.warning(
f"Process rank: {training_args.local_rank}, device: {training_args.device}, n_gpu: {training_args.n_gpu}"
+ f"distributed training: {bool(training_args.local_rank != -1)}, 16-bits training: {training_args.fp16}"
)
logger.info(f"Training/evaluation parameters {training_args}")
# Detecting last checkpoint.
last_checkpoint = None
if os.path.isdir(training_args.output_dir) and training_args.do_train and not training_args.overwrite_output_dir:
last_checkpoint = get_last_checkpoint(training_args.output_dir)
if last_checkpoint is None and len(os.listdir(training_args.output_dir)) > 0:
raise ValueError(
f"Output directory ({training_args.output_dir}) already exists and is not empty. "
"Use --overwrite_output_dir to overcome."
)
elif last_checkpoint is not None and training_args.resume_from_checkpoint is None:
logger.info(
f"Checkpoint detected, resuming training at {last_checkpoint}. To avoid this behavior, change "
"the `--output_dir` or add `--overwrite_output_dir` to train from scratch."
)
# Set seed before initializing model.
set_seed(training_args.seed)
# Get the datasets
# In distributed training, the load_dataset function guarantee that only one local process can concurrently
# download the dataset.
if data_args.dataset_name == "funsd":
# Downloading and loading a dataset from the hub.
dataset = load_dataset(
"nielsr/funsd-layoutlmv3",
data_args.dataset_config_name,
cache_dir=model_args.cache_dir,
use_auth_token=True if model_args.use_auth_token else None,
)
elif data_args.dataset_name == "cord":
# Downloading and loading a dataset from the hub.
dataset = load_dataset(
"nielsr/cord-layoutlmv3",
data_args.dataset_config_name,
cache_dir=model_args.cache_dir,
use_auth_token=True if model_args.use_auth_token else None,
)
else:
raise ValueError("This script only supports either FUNSD or CORD out-of-the-box.")
if training_args.do_train:
column_names = dataset["train"].column_names
features = dataset["train"].features
else:
column_names = dataset["test"].column_names
features = dataset["test"].features
image_column_name = "image"
text_column_name = "words" if "words" in column_names else "tokens"
boxes_column_name = "bboxes"
label_column_name = (
f"{data_args.task_name}_tags" if f"{data_args.task_name}_tags" in column_names else column_names[1]
)
remove_columns = column_names
# In the event the labels are not a `Sequence[ClassLabel]`, we will need to go through the dataset to get the
# unique labels.
def get_label_list(labels):
unique_labels = set()
for label in labels:
unique_labels = unique_labels | set(label)
label_list = list(unique_labels)
label_list.sort()
return label_list
# If the labels are of type ClassLabel, they are already integers and we have the map stored somewhere.
# Otherwise, we have to get the list of labels manually.
if isinstance(features[label_column_name].feature, ClassLabel):
label_list = features[label_column_name].feature.names
# No need to convert the labels since they are already ints.
id2label = {k: v for k, v in enumerate(label_list)}
label2id = {v: k for k, v in enumerate(label_list)}
else:
label_list = get_label_list(datasets["train"][label_column_name])
id2label = {k: v for k, v in enumerate(label_list)}
label2id = {v: k for k, v in enumerate(label_list)}
num_labels = len(label_list)
# Load pretrained model and processor
#
# Distributed training:
# The .from_pretrained methods guarantee that only one local process can concurrently
# download model & vocab.
config = AutoConfig.from_pretrained(
model_args.config_name if model_args.config_name else model_args.model_name_or_path,
num_labels=num_labels,
finetuning_task=data_args.task_name,
cache_dir=model_args.cache_dir,
revision=model_args.model_revision,
use_auth_token=True if model_args.use_auth_token else None,
)
processor = AutoProcessor.from_pretrained(
model_args.processor_name if model_args.processor_name else model_args.model_name_or_path,
cache_dir=model_args.cache_dir,
use_fast=True,
revision=model_args.model_revision,
use_auth_token=True if model_args.use_auth_token else None,
add_prefix_space=True,
apply_ocr=False,
)
model = AutoModelForTokenClassification.from_pretrained(
model_args.model_name_or_path,
from_tf=bool(".ckpt" in model_args.model_name_or_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,
)
# Set the correspondences label/ID inside the model config
model.config.label2id = label2id
model.config.id2label = id2label
# Preprocessing the dataset
# The processor does everything for us (prepare the image using LayoutLMv3FeatureExtractor
# and prepare the words, boxes and word-level labels using LayoutLMv3TokenizerFast)
def prepare_examples(examples):
images = examples[image_column_name]
words = examples[text_column_name]
boxes = examples[boxes_column_name]
word_labels = examples[label_column_name]
encoding = processor(
images,
words,
boxes=boxes,
word_labels=word_labels,
truncation=True,
padding="max_length",
max_length=data_args.max_seq_length,
)
return encoding
if training_args.do_train:
if "train" not in dataset:
raise ValueError("--do_train requires a train dataset")
train_dataset = dataset["train"]
if data_args.max_train_samples is not None:
train_dataset = train_dataset.select(range(data_args.max_train_samples))
with training_args.main_process_first(desc="train dataset map pre-processing"):
train_dataset = train_dataset.map(
prepare_examples,
batched=True,
remove_columns=remove_columns,
num_proc=data_args.preprocessing_num_workers,
load_from_cache_file=not data_args.overwrite_cache,
)
if training_args.do_eval:
validation_name = "test"
if validation_name not in dataset:
raise ValueError("--do_eval requires a validation dataset")
eval_dataset = dataset[validation_name]
if data_args.max_eval_samples is not None:
eval_dataset = eval_dataset.select(range(data_args.max_eval_samples))
with training_args.main_process_first(desc="validation dataset map pre-processing"):
eval_dataset = eval_dataset.map(
prepare_examples,
batched=True,
remove_columns=remove_columns,
num_proc=data_args.preprocessing_num_workers,
load_from_cache_file=not data_args.overwrite_cache,
)
if training_args.do_predict:
if "test" not in datasets:
raise ValueError("--do_predict requires a test dataset")
predict_dataset = datasets["test"]
if data_args.max_predict_samples is not None:
max_predict_samples = min(len(predict_dataset), data_args.max_predict_samples)
predict_dataset = predict_dataset.select(range(max_predict_samples))
with training_args.main_process_first(desc="prediction dataset map pre-processing"):
predict_dataset = predict_dataset.map(
prepare_examples,
batched=True,
remove_columns=remove_columns,
num_proc=data_args.preprocessing_num_workers,
load_from_cache_file=not data_args.overwrite_cache,
)
# Metrics
metric = load_metric("seqeval")
def compute_metrics(p):
predictions, labels = p
predictions = np.argmax(predictions, axis=2)
# Remove ignored index (special tokens)
true_predictions = [
[label_list[p] for (p, l) in zip(prediction, label) if l != -100]
for prediction, label in zip(predictions, labels)
]
true_labels = [
[label_list[l] for (p, l) in zip(prediction, label) if l != -100]
for prediction, label in zip(predictions, labels)
]
results = metric.compute(predictions=true_predictions, references=true_labels)
if data_args.return_entity_level_metrics:
# Unpack nested dictionaries
final_results = {}
for key, value in results.items():
if isinstance(value, dict):
for n, v in value.items():
final_results[f"{key}_{n}"] = v
else:
final_results[key] = value
return final_results
else:
return {
"precision": results["overall_precision"],
"recall": results["overall_recall"],
"f1": results["overall_f1"],
"accuracy": results["overall_accuracy"],
}
# Initialize our Trainer
trainer = Trainer(
model=model,
args=training_args,
train_dataset=train_dataset if training_args.do_train else None,
eval_dataset=eval_dataset if training_args.do_eval else None,
tokenizer=processor,
data_collator=default_data_collator,
compute_metrics=compute_metrics,
)
# Training
if training_args.do_train:
checkpoint = None
if training_args.resume_from_checkpoint is not None:
checkpoint = training_args.resume_from_checkpoint
elif last_checkpoint is not None:
checkpoint = last_checkpoint
train_result = trainer.train(resume_from_checkpoint=checkpoint)
metrics = train_result.metrics
trainer.save_model() # Saves the tokenizer too for easy upload
max_train_samples = (
data_args.max_train_samples if data_args.max_train_samples is not None else len(train_dataset)
)
metrics["train_samples"] = min(max_train_samples, len(train_dataset))
trainer.log_metrics("train", metrics)
trainer.save_metrics("train", metrics)
trainer.save_state()
# Evaluation
if training_args.do_eval:
logger.info("*** Evaluate ***")
metrics = trainer.evaluate()
max_eval_samples = data_args.max_eval_samples if data_args.max_eval_samples is not None else len(eval_dataset)
metrics["eval_samples"] = min(max_eval_samples, len(eval_dataset))
trainer.log_metrics("eval", metrics)
trainer.save_metrics("eval", metrics)
# Predict
if training_args.do_predict:
logger.info("*** Predict ***")
predictions, labels, metrics = trainer.predict(predict_dataset, metric_key_prefix="predict")
predictions = np.argmax(predictions, axis=2)
# Remove ignored index (special tokens)
true_predictions = [
[label_list[p] for (p, l) in zip(prediction, label) if l != -100]
for prediction, label in zip(predictions, labels)
]
trainer.log_metrics("predict", metrics)
trainer.save_metrics("predict", metrics)
# Save predictions
output_predictions_file = os.path.join(training_args.output_dir, "predictions.txt")
if trainer.is_world_process_zero():
with open(output_predictions_file, "w") as writer:
for prediction in true_predictions:
writer.write(" ".join(prediction) + "\n")
kwargs = {"finetuned_from": model_args.model_name_or_path, "tasks": "token-classification"}
if data_args.dataset_name is not None:
kwargs["dataset_tags"] = data_args.dataset_name
if data_args.dataset_config_name is not None:
kwargs["dataset_args"] = data_args.dataset_config_name
kwargs["dataset"] = f"{data_args.dataset_name} {data_args.dataset_config_name}"
else:
kwargs["dataset"] = data_args.dataset_name
if training_args.push_to_hub:
trainer.push_to_hub(**kwargs)
else:
trainer.create_model_card(**kwargs)
def _mp_fn(index):
# For xla_spawn (TPUs)
main()
if __name__ == "__main__":
main()