mirror of
https://github.com/huggingface/transformers.git
synced 2025-07-14 18:18:24 +06:00
9aeacb58ba
* [WIP] SP tokenizers * fixing tests for T5 * WIP tokenizers * serialization * update T5 * WIP T5 tokenization * slow to fast conversion script * Refactoring to move tokenzier implementations inside transformers * Adding gpt - refactoring - quality * WIP adding several tokenizers to the fast world * WIP Roberta - moving implementations * update to dev4 switch file loading to in-memory loading * Updating and fixing * advancing on the tokenizers - updating do_lower_case * style and quality * moving forward with tokenizers conversion and tests * MBart, T5 * dumping the fast version of transformer XL * Adding to autotokenizers + style/quality * update init and space_between_special_tokens * style and quality * bump up tokenizers version * add protobuf * fix pickle Bert JP with Mecab * fix newly added tokenizers * style and quality * fix bert japanese * fix funnel * limite tokenizer warning to one occurence * clean up file * fix new tokenizers * fast tokenizers deep tests * WIP adding all the special fast tests on the new fast tokenizers * quick fix * adding more fast tokenizers in the fast tests * all tokenizers in fast version tested * Adding BertGenerationFast * bump up setup.py for CI * remove BertGenerationFast (too early) * bump up tokenizers version * Clean old docstrings * Typo * Update following Lysandre comments Co-authored-by: Sylvain Gugger <sylvain.gugger@gmail.com>
87 lines
3.4 KiB
Python
87 lines
3.4 KiB
Python
# coding=utf-8
|
|
# Copyright 2020 Huggingface
|
|
#
|
|
# 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.
|
|
|
|
|
|
from transformers.testing_utils import slow
|
|
from transformers.tokenization_dpr import (
|
|
DPRContextEncoderTokenizer,
|
|
DPRContextEncoderTokenizerFast,
|
|
DPRQuestionEncoderTokenizer,
|
|
DPRQuestionEncoderTokenizerFast,
|
|
DPRReaderOutput,
|
|
DPRReaderTokenizer,
|
|
DPRReaderTokenizerFast,
|
|
)
|
|
from transformers.tokenization_utils_base import BatchEncoding
|
|
|
|
from .test_tokenization_bert import BertTokenizationTest
|
|
|
|
|
|
class DPRContextEncoderTokenizationTest(BertTokenizationTest):
|
|
|
|
tokenizer_class = DPRContextEncoderTokenizer
|
|
rust_tokenizer_class = DPRContextEncoderTokenizerFast
|
|
test_rust_tokenizer = True
|
|
|
|
|
|
class DPRQuestionEncoderTokenizationTest(BertTokenizationTest):
|
|
|
|
tokenizer_class = DPRQuestionEncoderTokenizer
|
|
rust_tokenizer_class = DPRQuestionEncoderTokenizerFast
|
|
test_rust_tokenizer = True
|
|
|
|
|
|
class DPRReaderTokenizationTest(BertTokenizationTest):
|
|
|
|
tokenizer_class = DPRReaderTokenizer
|
|
rust_tokenizer_class = DPRReaderTokenizerFast
|
|
test_rust_tokenizer = True
|
|
|
|
@slow
|
|
def test_decode_best_spans(self):
|
|
tokenizer = self.tokenizer_class.from_pretrained("bert-base-uncased")
|
|
|
|
text_1 = tokenizer.encode("question sequence", add_special_tokens=False)
|
|
text_2 = tokenizer.encode("title sequence", add_special_tokens=False)
|
|
text_3 = tokenizer.encode("text sequence " * 4, add_special_tokens=False)
|
|
input_ids = [[101] + text_1 + [102] + text_2 + [102] + text_3]
|
|
reader_input = BatchEncoding({"input_ids": input_ids})
|
|
|
|
start_logits = [[0] * len(input_ids[0])]
|
|
end_logits = [[0] * len(input_ids[0])]
|
|
relevance_logits = [0]
|
|
reader_output = DPRReaderOutput(start_logits, end_logits, relevance_logits)
|
|
|
|
start_index, end_index = 8, 9
|
|
start_logits[0][start_index] = 10
|
|
end_logits[0][end_index] = 10
|
|
predicted_spans = tokenizer.decode_best_spans(reader_input, reader_output)
|
|
self.assertEqual(predicted_spans[0].start_index, start_index)
|
|
self.assertEqual(predicted_spans[0].end_index, end_index)
|
|
self.assertEqual(predicted_spans[0].doc_id, 0)
|
|
|
|
@slow
|
|
def test_call(self):
|
|
tokenizer = self.tokenizer_class.from_pretrained("bert-base-uncased")
|
|
|
|
text_1 = tokenizer.encode("question sequence", add_special_tokens=False)
|
|
text_2 = tokenizer.encode("title sequence", add_special_tokens=False)
|
|
text_3 = tokenizer.encode("text sequence", add_special_tokens=False)
|
|
expected_input_ids = [101] + text_1 + [102] + text_2 + [102] + text_3
|
|
encoded_input = tokenizer(questions=["question sequence"], titles=["title sequence"], texts=["text sequence"])
|
|
self.assertIn("input_ids", encoded_input)
|
|
self.assertIn("attention_mask", encoded_input)
|
|
self.assertListEqual(encoded_input["input_ids"][0], expected_input_ids)
|