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

Model templates encoder only (#8509)

Browse Source 
* Model templates

* TensorFlow

* Remove pooler

* CI

* Tokenizer + Refactoring

* Encoder-Decoder

* Let's go testing

* Encoder-Decoder in TF

* Let's go testing in TF

* Documentation

* README

* Fixes

* Better names

* Style

* Update docs

* Choose to skip either TF or PT

* Code quality fixes

* Add to testing suite

* Update file path

* Cookiecutter path

* Update `transformers` path

* Handle rebasing

* Remove seq2seq from model templates

* Remove s2s config

* Apply Sylvain and Patrick comments

* Apply suggestions from code review

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* Last fixes from code review

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>
This commit is contained in:
Lysandre Debut
2020-11-13 11:59:30 -05:00
committed by GitHub
parent 42e2d02e44
commit 826f04576f
29 changed files with 3315 additions and 1977 deletions
Download Patch File Download Diff File Expand all files Collapse all files

10
templates/adding_a_new_model/tests/encoder-bert-tokenizer.json Normal file
View File

@@ -0,0 +1,10 @@
{
"modelname": "EncoderBERT",
"uppercase_modelname": "ENCODER_BERT",
"lowercase_modelname": "encoder_bert",
"camelcase_modelname": "EncoderBert",
"authors": "The HuggingFace Team",
"checkpoint_identifier": "brand-new-bert-base-cased",
"tokenizer_type": "Based on BERT",
"generate_tensorflow_and_pytorch": "PyTorch & TensorFlow"
}

10
templates/adding_a_new_model/tests/pt-encoder-bert-tokenizer.json Normal file
View File

@@ -0,0 +1,10 @@
{
"modelname": "PTEncoderBERT",
"uppercase_modelname": "PT_ENCODER_BERT",
"lowercase_modelname": "pt_encoder_bert",
"camelcase_modelname": "PtEncoderBert",
"authors": "The HuggingFace Team",
"checkpoint_identifier": "brand-new-bert-base-cased",
"tokenizer_type": "Based on BERT",
"generate_tensorflow_and_pytorch": "PyTorch"
}

10
templates/adding_a_new_model/tests/standalone.json Normal file
View File

@@ -0,0 +1,10 @@
{
"modelname": "BIEncoderBERT",
"uppercase_modelname": "BI_ENCODER_BERT",
"lowercase_modelname": "bi_encoder_bert",
"camelcase_modelname": "BiEncoderBert",
"authors": "The HuggingFace Team",
"checkpoint_identifier": "bi-brand-new-bert-base-cased",
"tokenizer_type": "Standalone",
"generate_tensorflow_and_pytorch": "PyTorch & TensorFlow"
}

257
templates/adding_a_new_model/tests/test_modeling_tf_xxx.py
View File

@@ -1,257 +0,0 @@
# coding=utf-8
# Copyright 2018 XXX Authors.
#
# 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.
import unittest
from transformers import XxxConfig, is_tf_available
from transformers.testing_utils import CACHE_DIR, require_tf, slow
from .test_configuration_common import ConfigTester
from .test_modeling_tf_common import TFModelTesterMixin, ids_tensor
if is_tf_available():
import tensorflow as tf
from transformers.modeling_tf_xxx import (
TFXxxForMaskedLM,
TFXxxForMultipleChoice,
TFXxxForQuestionAnswering,
TFXxxForSequenceClassification,
TFXxxForTokenClassification,
TFXxxModel,
)
@require_tf
class TFXxxModelTest(TFModelTesterMixin, unittest.TestCase):
all_model_classes = (
(
TFXxxModel,
TFXxxForMaskedLM,
TFXxxForMultipleChoice,
TFXxxForQuestionAnswering,
TFXxxForSequenceClassification,
TFXxxForTokenClassification,
)
if is_tf_available()
else ()
)
class TFXxxModelTester(object):
def __init__(
self,
parent,
batch_size=13,
seq_length=7,
is_training=True,
use_input_mask=True,
use_token_type_ids=True,
use_labels=True,
vocab_size=99,
hidden_size=32,
num_hidden_layers=5,
num_attention_heads=4,
intermediate_size=37,
hidden_act="gelu",
hidden_dropout_prob=0.1,
attention_probs_dropout_prob=0.1,
max_position_embeddings=512,
type_vocab_size=16,
type_sequence_label_size=2,
initializer_range=0.02,
num_labels=3,
num_choices=4,
scope=None,
):
self.parent = parent
self.batch_size = batch_size
self.seq_length = seq_length
self.is_training = is_training
self.use_input_mask = use_input_mask
self.use_token_type_ids = use_token_type_ids
self.use_labels = use_labels
self.vocab_size = vocab_size
self.hidden_size = hidden_size
self.num_hidden_layers = num_hidden_layers
self.num_attention_heads = num_attention_heads
self.intermediate_size = intermediate_size
self.hidden_act = hidden_act
self.hidden_dropout_prob = hidden_dropout_prob
self.attention_probs_dropout_prob = attention_probs_dropout_prob
self.max_position_embeddings = max_position_embeddings
self.type_vocab_size = type_vocab_size
self.type_sequence_label_size = type_sequence_label_size
self.initializer_range = initializer_range
self.num_labels = num_labels
self.num_choices = num_choices
self.scope = scope
def prepare_config_and_inputs(self):
input_ids = ids_tensor([self.batch_size, self.seq_length], self.vocab_size)
input_mask = None
if self.use_input_mask:
input_mask = ids_tensor([self.batch_size, self.seq_length], vocab_size=2)
token_type_ids = None
if self.use_token_type_ids:
token_type_ids = ids_tensor([self.batch_size, self.seq_length], self.type_vocab_size)
sequence_labels = None
token_labels = None
choice_labels = None
if self.use_labels:
sequence_labels = ids_tensor([self.batch_size], self.type_sequence_label_size)
token_labels = ids_tensor([self.batch_size, self.seq_length], self.num_labels)
choice_labels = ids_tensor([self.batch_size], self.num_choices)
config = XxxConfig(
vocab_size=self.vocab_size,
hidden_size=self.hidden_size,
num_hidden_layers=self.num_hidden_layers,
num_attention_heads=self.num_attention_heads,
intermediate_size=self.intermediate_size,
hidden_act=self.hidden_act,
hidden_dropout_prob=self.hidden_dropout_prob,
attention_probs_dropout_prob=self.attention_probs_dropout_prob,
max_position_embeddings=self.max_position_embeddings,
type_vocab_size=self.type_vocab_size,
initializer_range=self.initializer_range,
return_dict=True,
)
return config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
def create_and_check_model(
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
):
model = TFXxxModel(config=config)
inputs = {"input_ids": input_ids, "attention_mask": input_mask, "token_type_ids": token_type_ids}
result = model(inputs)
inputs = [input_ids, input_mask]
result = model(inputs)
result = model(input_ids)
self.parent.assertEqual(
result.last_hidden_state.shape, (self.batch_size, self.seq_length, self.hidden_size)
)
self.parent.assertEqual(result.pooler_output.shape, (self.batch_size, self.hidden_size))
def create_and_check_for_masked_lm(
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
):
model = TFXxxForMaskedLM(config=config)
inputs = {"input_ids": input_ids, "attention_mask": input_mask, "token_type_ids": token_type_ids}
result = model(inputs)
self.parent.assertEqual(result.logits.shape, (self.batch_size, self.seq_length, self.vocab_size))
def create_and_check_for_sequence_classification(
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
):
config.num_labels = self.num_labels
model = TFXxxForSequenceClassification(config=config)
inputs = {"input_ids": input_ids, "attention_mask": input_mask, "token_type_ids": token_type_ids}
result = model(inputs)
self.parent.assertEqual(result.logits.shape, (self.batch_size, self.num_labels))
def create_and_check_for_multiple_choice(
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
):
config.num_choices = self.num_choices
model = TFXxxForMultipleChoice(config=config)
multiple_choice_inputs_ids = tf.tile(tf.expand_dims(input_ids, 1), (1, self.num_choices, 1))
multiple_choice_input_mask = tf.tile(tf.expand_dims(input_mask, 1), (1, self.num_choices, 1))
multiple_choice_token_type_ids = tf.tile(tf.expand_dims(token_type_ids, 1), (1, self.num_choices, 1))
inputs = {
"input_ids": multiple_choice_inputs_ids,
"attention_mask": multiple_choice_input_mask,
"token_type_ids": multiple_choice_token_type_ids,
}
result = model(inputs)
self.parent.assertEqual(result.logits.shape, (self.batch_size, self.num_choices))
def create_and_check_for_token_classification(
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
):
config.num_labels = self.num_labels
model = TFXxxForTokenClassification(config=config)
inputs = {"input_ids": input_ids, "attention_mask": input_mask, "token_type_ids": token_type_ids}
result = model(inputs)
self.parent.assertEqual(result.logits.shape, (self.batch_size, self.seq_length, self.num_labels))
def create_and_check_for_question_answering(
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
):
model = TFXxxForQuestionAnswering(config=config)
inputs = {"input_ids": input_ids, "attention_mask": input_mask, "token_type_ids": token_type_ids}
result = model(inputs)
self.parent.assertEqual(result.start_logits.shape, (self.batch_size, self.seq_length))
self.parent.assertEqual(result.end_logits.shape, (self.batch_size, self.seq_length))
def prepare_config_and_inputs_for_common(self):
config_and_inputs = self.prepare_config_and_inputs()
(
config,
input_ids,
token_type_ids,
input_mask,
sequence_labels,
token_labels,
choice_labels,
) = config_and_inputs
inputs_dict = {"input_ids": input_ids, "token_type_ids": token_type_ids, "attention_mask": input_mask}
return config, inputs_dict
def setUp(self):
self.model_tester = TFXxxModelTest.TFXxxModelTester(self)
self.config_tester = ConfigTester(self, config_class=XxxConfig, hidden_size=37)
def test_config(self):
self.config_tester.run_common_tests()
def test_model(self):
config_and_inputs = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*config_and_inputs)
def test_for_masked_lm(self):
config_and_inputs = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_masked_lm(*config_and_inputs)
def test_for_question_answering(self):
config_and_inputs = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_question_answering(*config_and_inputs)
def test_for_sequence_classification(self):
config_and_inputs = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_sequence_classification(*config_and_inputs)
def test_for_token_classification(self):
config_and_inputs = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_token_classification(*config_and_inputs)
def test_for_multiple_choice(self):
config_and_inputs = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_multiple_choice(*config_and_inputs)
@slow
def test_model_from_pretrained(self):
for model_name in ["xxx-base-uncased"]:
model = TFXxxModel.from_pretrained(model_name, cache_dir=CACHE_DIR)
self.assertIsNotNone(model)

307
templates/adding_a_new_model/tests/test_modeling_xxx.py
View File

@@ -1,307 +0,0 @@
# coding=utf-8
# Copyright 2018 XXX Authors.
#
# 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.
import unittest
from transformers import is_torch_available
from transformers.testing_utils import require_torch, require_torch_gpu, slow, torch_device
from .test_configuration_common import ConfigTester
from .test_modeling_common import ModelTesterMixin, ids_tensor
if is_torch_available():
from transformers import (
AutoModelForMaskedLM,
AutoTokenizer,
XxxConfig,
XxxForMaskedLM,
XxxForMultipleChoice,
XxxForQuestionAnswering,
XxxForSequenceClassification,
XxxForTokenClassification,
XxxModel,
)
from transformers.file_utils import cached_property
#
class XxxModelTester:
"""You can also import this e.g from .test_modeling_bart import BartModelTester """
def __init__(
self,
parent,
batch_size=13,
seq_length=7,
is_training=True,
use_input_mask=True,
use_token_type_ids=True,
use_labels=True,
vocab_size=99,
hidden_size=32,
num_hidden_layers=5,
num_attention_heads=4,
intermediate_size=37,
hidden_act="gelu",
hidden_dropout_prob=0.1,
attention_probs_dropout_prob=0.1,
max_position_embeddings=512,
type_vocab_size=16,
type_sequence_label_size=2,
initializer_range=0.02,
num_labels=3,
num_choices=4,
scope=None,
):
self.parent = parent
self.batch_size = batch_size
self.seq_length = seq_length
self.is_training = is_training
self.use_input_mask = use_input_mask
self.use_token_type_ids = use_token_type_ids
self.use_labels = use_labels
self.vocab_size = vocab_size
self.hidden_size = hidden_size
self.num_hidden_layers = num_hidden_layers
self.num_attention_heads = num_attention_heads
self.intermediate_size = intermediate_size
self.hidden_act = hidden_act
self.hidden_dropout_prob = hidden_dropout_prob
self.attention_probs_dropout_prob = attention_probs_dropout_prob
self.max_position_embeddings = max_position_embeddings
self.type_vocab_size = type_vocab_size
self.type_sequence_label_size = type_sequence_label_size
self.initializer_range = initializer_range
self.num_labels = num_labels
self.num_choices = num_choices
self.scope = scope
def prepare_config_and_inputs(self):
input_ids = ids_tensor([self.batch_size, self.seq_length], self.vocab_size)
input_mask = None
if self.use_input_mask:
input_mask = ids_tensor([self.batch_size, self.seq_length], vocab_size=2)
token_type_ids = None
if self.use_token_type_ids:
token_type_ids = ids_tensor([self.batch_size, self.seq_length], self.type_vocab_size)
sequence_labels = None
token_labels = None
choice_labels = None
if self.use_labels:
sequence_labels = ids_tensor([self.batch_size], self.type_sequence_label_size)
token_labels = ids_tensor([self.batch_size, self.seq_length], self.num_labels)
choice_labels = ids_tensor([self.batch_size], self.num_choices)
config = XxxConfig(
vocab_size=self.vocab_size,
hidden_size=self.hidden_size,
num_hidden_layers=self.num_hidden_layers,
num_attention_heads=self.num_attention_heads,
intermediate_size=self.intermediate_size,
hidden_act=self.hidden_act,
hidden_dropout_prob=self.hidden_dropout_prob,
attention_probs_dropout_prob=self.attention_probs_dropout_prob,
max_position_embeddings=self.max_position_embeddings,
type_vocab_size=self.type_vocab_size,
initializer_range=self.initializer_range,
return_dict=True,
)
return config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
def create_and_check_model(
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
):
model = XxxModel(config=config)
model.to(torch_device)
model.eval()
result = model(input_ids, attention_mask=input_mask, token_type_ids=token_type_ids)
result = model(input_ids, token_type_ids=token_type_ids)
result = model(input_ids)
self.parent.assertEqual(result.last_hidden_state.shape, (self.batch_size, self.seq_length, self.hidden_size))
self.parent.assertEqual(result.pooler_output.shape, (self.batch_size, self.hidden_size))
def create_and_check_for_masked_lm(
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
):
model = XxxForMaskedLM(config=config)
model.to(torch_device)
model.eval()
result = model(input_ids, attention_mask=input_mask, token_type_ids=token_type_ids, labels=token_labels)
self.parent.assertEqual(result.logits.shape, (self.batch_size, self.seq_length, self.vocab_size))
def create_and_check_for_question_answering(
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
):
model = XxxForQuestionAnswering(config=config)
model.to(torch_device)
model.eval()
result = model(
input_ids,
attention_mask=input_mask,
token_type_ids=token_type_ids,
start_positions=sequence_labels,
end_positions=sequence_labels,
)
self.parent.assertEqual(result.start_logits.shape, (self.batch_size, self.seq_length))
self.parent.assertEqual(result.end_logits.shape, (self.batch_size, self.seq_length))
def create_and_check_for_sequence_classification(
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
):
config.num_labels = self.num_labels
model = XxxForSequenceClassification(config)
model.to(torch_device)
model.eval()
result = model(input_ids, attention_mask=input_mask, token_type_ids=token_type_ids, labels=sequence_labels)
self.parent.assertEqual(result.logits.shape, (self.batch_size, self.num_labels))
def create_and_check_for_token_classification(
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
):
config.num_labels = self.num_labels
model = XxxForTokenClassification(config=config)
model.to(torch_device)
model.eval()
result = model(input_ids, attention_mask=input_mask, token_type_ids=token_type_ids, labels=token_labels)
self.parent.assertEqual(result.logits.shape, (self.batch_size, self.seq_length, self.num_labels))
def create_and_check_for_multiple_choice(
self, config, input_ids, token_type_ids, input_mask, sequence_labels, token_labels, choice_labels
):
config.num_choices = self.num_choices
model = XxxForMultipleChoice(config=config)
model.to(torch_device)
model.eval()
multiple_choice_inputs_ids = input_ids.unsqueeze(1).expand(-1, self.num_choices, -1).contiguous()
multiple_choice_token_type_ids = token_type_ids.unsqueeze(1).expand(-1, self.num_choices, -1).contiguous()
multiple_choice_input_mask = input_mask.unsqueeze(1).expand(-1, self.num_choices, -1).contiguous()
result = model(
multiple_choice_inputs_ids,
attention_mask=multiple_choice_input_mask,
token_type_ids=multiple_choice_token_type_ids,
labels=choice_labels,
)
self.parent.assertEqual(result.logits.shape, (self.batch_size, self.num_choices))
def prepare_config_and_inputs_for_common(self):
config_and_inputs = self.prepare_config_and_inputs()
(
config,
input_ids,
token_type_ids,
input_mask,
sequence_labels,
token_labels,
choice_labels,
) = config_and_inputs
inputs_dict = {"input_ids": input_ids, "token_type_ids": token_type_ids, "attention_mask": input_mask}
return config, inputs_dict
@require_torch
class XxxModelTest(ModelTesterMixin, unittest.TestCase):
all_model_classes = (
(XxxModel, XxxForMaskedLM, XxxForQuestionAnswering, XxxForSequenceClassification, XxxForTokenClassification)
if is_torch_available()
else ()
)
def setUp(self):
self.model_tester = XxxModelTester(self)
self.config_tester = ConfigTester(self, config_class=XxxConfig, hidden_size=37)
def test_config(self):
self.config_tester.run_common_tests()
def test_model(self):
config_and_inputs = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_model(*config_and_inputs)
def test_for_masked_lm(self):
config_and_inputs = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_masked_lm(*config_and_inputs)
def test_for_question_answering(self):
config_and_inputs = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_question_answering(*config_and_inputs)
def test_for_sequence_classification(self):
config_and_inputs = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_sequence_classification(*config_and_inputs)
def test_for_token_classification(self):
config_and_inputs = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_for_token_classification(*config_and_inputs)
def test_for_multiple_choice(self):
config_and_inputs = self.model_tester.prepare_config_and_inputs()
self.model_tester.create_and_check_electra_for_multiple_choice(*config_and_inputs)
@slow
def test_lm_outputs_same_as_reference_model(self):
"""Write something that could help someone fixing this here."""
checkpoint_path = "XXX/bart-large"
model = self.big_model
tokenizer = AutoTokenizer.from_pretrained(
checkpoint_path
) # same with AutoTokenizer (see tokenization_auto.py). This is not mandatory
# MODIFY THIS DEPENDING ON YOUR MODELS RELEVANT TASK.
batch = tokenizer(["I went to the <mask> yesterday"]).to(torch_device)
desired_mask_result = tokenizer.decode("store") # update this
logits = model(**batch).logits
masked_index = (batch.input_ids == self.tokenizer.mask_token_id).nonzero()
assert model.num_parameters() == 175e9 # a joke
mask_entry_logits = logits[0, masked_index.item(), :]
probs = mask_entry_logits.softmax(dim=0)
_, predictions = probs.topk(1)
self.assertEqual(tokenizer.decode(predictions), desired_mask_result)
@cached_property
def big_model(self):
"""Cached property means this code will only be executed once."""
checkpoint_path = "XXX/bart-large"
model = AutoModelForMaskedLM.from_pretrained(checkpoint_path).to(
torch_device
) # test whether AutoModel can determine your model_class from checkpoint name
if torch_device == "cuda":
model.half()
# optional: do more testing! This will save you time later!
@slow
def test_that_XXX_can_be_used_in_a_pipeline(self):
"""We can use self.big_model here without calling __init__ again."""
pass
def test_XXX_loss_doesnt_change_if_you_add_padding(self):
pass
def test_XXX_bad_args(self):
pass
def test_XXX_backward_pass_reduces_loss(self):
"""Test loss/gradients same as reference implementation, for example."""
pass
@require_torch_gpu
def test_large_inputs_in_fp16_dont_cause_overflow(self):
pass

68
templates/adding_a_new_model/tests/test_tokenization_xxx.py
View File

@@ -1,68 +0,0 @@
# coding=utf-8
# Copyright 2018 XXX Authors.
#
# 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.
import os
import unittest
from transformers.tokenization_bert import VOCAB_FILES_NAMES, XxxTokenizer
from .test_tokenization_common import TokenizerTesterMixin
class XxxTokenizationTest(TokenizerTesterMixin, unittest.TestCase):
tokenizer_class = XxxTokenizer
def setUp(self):
super().setUp()
vocab_tokens = [
"[UNK]",
"[CLS]",
"[SEP]",
"want",
"##want",
"##ed",
"wa",
"un",
"runn",
"##ing",
",",
"low",
"lowest",
]
self.vocab_file = os.path.join(self.tmpdirname, VOCAB_FILES_NAMES["vocab_file"])
with open(self.vocab_file, "w", encoding="utf-8") as vocab_writer:
vocab_writer.write("".join([x + "\n" for x in vocab_tokens]))
def get_tokenizer(self, **kwargs):
return XxxTokenizer.from_pretrained(self.tmpdirname, **kwargs)
def get_input_output_texts(self, tokenizer):
input_text = "UNwant\u00E9d,running"
output_text = "unwanted, running"
return input_text, output_text
def test_full_tokenizer(self):
tokenizer = self.tokenizer_class(self.vocab_file)
tokens = tokenizer.tokenize("UNwant\u00E9d,running")
self.assertListEqual(tokens, ["un", "##want", "##ed", ",", "runn", "##ing"])
self.assertListEqual(tokenizer.convert_tokens_to_ids(tokens), [7, 4, 5, 10, 8, 9])
def test_special_tokens_as_you_expect(self):
"""If you are training a seq2seq model that expects a decoder_prefix token make sure it is prepended to decoder_input_ids """
pass

10
templates/adding_a_new_model/tests/tf-encoder-bert-tokenizer.json Normal file
View File

@@ -0,0 +1,10 @@
{
"modelname": "TFEncoderBERT",
"uppercase_modelname": "TF_ENCODER_BERT",
"lowercase_modelname": "tf_encoder_bert",
"camelcase_modelname": "TfEncoderBert",
"authors": "The HuggingFace Team",
"checkpoint_identifier": "brand-new-bert-base-cased",
"tokenizer_type": "Based on BERT",
"generate_tensorflow_and_pytorch": "TensorFlow"
}