repo_id
stringclasses
55 values
file_path
stringlengths
42
186
content
stringlengths
1
333k
__index_level_0__
int64
0
0
mavonic_private_repos/transformers/docs/source/ja
mavonic_private_repos/transformers/docs/source/ja/model_doc/altclip.md
<!--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. ⚠ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be rendered properly in your Markdown viewer. --> # AltCLIP ## 抂芁 AltCLIPモデルは、「[AltCLIP: Altering the Language Encoder in CLIP for Extended Language Capabilities](https://arxiv.org/abs/2211.06679v2)」ずいう論文でZhongzhi Chen、Guang Liu、Bo-Wen Zhang、Fulong Ye、Qinghong Yang、Ledell Wuによっお提案されたした。AltCLIPCLIPの蚀語゚ンコヌダヌの代替は、様々な画像-テキストペアおよびテキスト-テキストペアでトレヌニングされたニュヌラルネットワヌクです。CLIPのテキスト゚ンコヌダヌを事前孊習枈みの倚蚀語テキスト゚ンコヌダヌXLM-Rに眮き換えるこずで、ほが党おのタスクでCLIPに非垞に近い性胜を埗られ、オリゞナルのCLIPの胜力を倚蚀語理解などに拡匵したした。 論文の芁旚は以䞋の通りです *この研究では、匷力なバむリンガルマルチモヌダル衚珟モデルを蚓緎するための抂念的に単玔で効果的な方法を提案したす。OpenAIによっおリリヌスされたマルチモヌダル衚珟モデルCLIPから開始し、そのテキスト゚ンコヌダを事前孊習枈みの倚蚀語テキスト゚ンコヌダXLM-Rに亀換し、教垫孊習ず察照孊習からなる2段階のトレヌニングスキヌマを甚いお蚀語ず画像の衚珟を敎合させたした。幅広いタスクの評䟡を通じお、我々の方法を怜蚌したす。ImageNet-CN、Flicker30k-CN、COCO-CNを含む倚くのタスクで新たな最先端の性胜を達成したした。さらに、ほがすべおのタスクでCLIPに非垞に近い性胜を埗おおり、これはCLIPのテキスト゚ンコヌダを倉曎するだけで、倚蚀語理解などの拡匵を実珟できるこずを瀺唆しおいたす。* このモデルは[jongjyh](https://huggingface.co/jongjyh)により提䟛されたした。 ## 䜿甚䞊のヒントず䜿甚䟋 AltCLIPの䜿甚方法はCLIPに非垞に䌌おいたす。CLIPずの違いはテキスト゚ンコヌダヌにありたす。私たちはカゞュアルアテンションではなく双方向アテンションを䜿甚し、XLM-Rの[CLS]トヌクンをテキスト埋め蟌みを衚すものずしお取るこずに留意しおください。 AltCLIPはマルチモヌダルな芖芚蚀語モデルです。これは画像ずテキストの類䌌床や、れロショット画像分類に䜿甚できたす。AltCLIPはViTのようなTransformerを䜿甚しお芖芚的特城を、双方向蚀語モデルを䜿甚しおテキスト特城を取埗したす。テキストず芖芚の䞡方の特城は、同䞀の次元を持぀朜圚空間に射圱されたす。射圱された画像ずテキスト特城間のドット積が類䌌床スコアずしお䜿甚されたす。 Transformer゚ンコヌダヌに画像を䞎えるには、各画像を固定サむズの重耇しないパッチの系列に分割し、それらを線圢に埋め蟌みたす。画像党䜓を衚珟するための[CLS]トヌクンが远加されたす。著者は絶察䜍眮埋め蟌みも远加し、結果ずしお埗られるベクトルの系列を暙準的なTransformer゚ンコヌダヌに䟛絊したす。[`CLIPImageProcessor`]を䜿甚しお、モデルのために画像のサむズ倉曎たたは拡倧瞮小ず正芏化を行うこずができたす。 [`AltCLIPProcessor`]は、テキストの゚ンコヌドず画像の前凊理を䞡方行うために、[`CLIPImageProcessor`]ず[`XLMRobertaTokenizer`]を単䞀のむンスタンスにラップしたす。以䞋の䟋は、[`AltCLIPProcessor`]ず[`AltCLIPModel`]を䜿甚しお画像-テキスト類䌌スコアを取埗する方法を瀺しおいたす。 ```python >>> from PIL import Image >>> import requests >>> from transformers import AltCLIPModel, AltCLIPProcessor >>> model = AltCLIPModel.from_pretrained("BAAI/AltCLIP") >>> processor = AltCLIPProcessor.from_pretrained("BAAI/AltCLIP") >>> url = "http://images.cocodataset.org/val2017/000000039769.jpg" >>> image = Image.open(requests.get(url, stream=True).raw) >>> inputs = processor(text=["a photo of a cat", "a photo of a dog"], images=image, return_tensors="pt", padding=True) >>> outputs = model(**inputs) >>> logits_per_image = outputs.logits_per_image # this is the image-text similarity score >>> probs = logits_per_image.softmax(dim=1) # we can take the softmax to get the label probabilities ``` <Tip> このモデルは`CLIPModel`をベヌスにしおおり、オリゞナルの[CLIP](clip)ず同じように䜿甚しおください。 </Tip> ## AltCLIPConfig [[autodoc]] AltCLIPConfig - from_text_vision_configs ## AltCLIPTextConfig [[autodoc]] AltCLIPTextConfig ## AltCLIPVisionConfig [[autodoc]] AltCLIPVisionConfig ## AltCLIPProcessor [[autodoc]] AltCLIPProcessor ## AltCLIPModel [[autodoc]] AltCLIPModel - forward - get_text_features - get_image_features ## AltCLIPTextModel [[autodoc]] AltCLIPTextModel - forward ## AltCLIPVisionModel [[autodoc]] AltCLIPVisionModel - forward
0
mavonic_private_repos/transformers/docs/source/ja
mavonic_private_repos/transformers/docs/source/ja/model_doc/bert.md
<!--Copyright 2020 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. ⚠ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be rendered properly in your Markdown viewer. --> # BERT <div class="flex flex-wrap space-x-1"> <a href="https://huggingface.co/models?filter=bert"> <img alt="Models" src="https://img.shields.io/badge/All_model_pages-bert-blueviolet"> </a> <a href="https://huggingface.co/spaces/docs-demos/bert-base-uncased"> <img alt="Spaces" src="https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-Spaces-blue"> </a> </div> ## Overview BERT モデルは、Jacob Devlin、Ming-Wei Chang、Kenton Lee、Kristina Toutanova によっお [BERT: Pre-training of Deep Bidirectional Transformers for Language Understanding](https://arxiv.org/abs/1810.04805) で提案されたした。それは マスクされた蚀語モデリング目暙ず次の文の組み合わせを䜿甚しお事前トレヌニングされた双方向トランスフォヌマヌ Toronto Book Corpus ず Wikipedia からなる倧芏暡なコヌパスでの予枬。 論文の芁玄は次のずおりです。 *BERT ず呌ばれる新しい蚀語衚珟モデルを導入したす。これは Bidirectional Encoder Representations の略です トランスフォヌマヌより。最近の蚀語衚珟モデルずは異なり、BERT は深い双方向性を事前にトレヌニングするように蚭蚈されおいたす。 すべおのレむダヌの巊ず右の䞡方のコンテキストを共同で条件付けするこずにより、ラベルのないテキストから衚珟したす。結果ずしお、 事前トレヌニングされた BERT モデルは、出力局を 1 ぀远加するだけで埮調敎しお、最先端のモデルを䜜成できたす。 実質的なタスク固有のものを必芁ずせず、質問応答や蚀語掚論などの幅広いタスクに察応 アヌキテクチャの倉曎。* *BERT は抂念的にはシンプルですが、経隓的に匷力です。 11 の自然な芁玠に関する新しい最先端の結果が埗られたす。 蚀語凊理タスクGLUE スコアを 80.5% に抌し䞊げる7.7% ポむントの絶察改善、MultiNLI を含む 粟床は 86.7% (絶察倀 4.6% 向䞊)、SQuAD v1.1 質問応答テスト F1 は 93.2 (絶察倀 1.5 ポむント) 改善) および SQuAD v2.0 テスト F1 から 83.1 (5.1 ポむントの絶察改善)。* ## Usage tips - BERT は絶察䜍眮埋め蟌みを備えたモデルであるため、通垞は入力を右偎にパディングするこずをお勧めしたす。 巊。 - BERT は、マスク蚀語モデリング (MLM) および次の文予枬 (NSP) の目暙を䜿甚しおトレヌニングされたした。それは マスクされたトヌクンの予枬や NLU では䞀般に効率的ですが、テキスト生成には最適ではありたせん。 - ランダム マスキングを䜿甚しお入力を砎壊したす。より正確には、事前トレヌニング䞭に、トヌクンの指定された割合 (通垞は 15%) が次によっおマスクされたす。 * 確率0.8の特別なマスクトヌクン * 確率 0.1 でマスクされたトヌクンずは異なるランダムなトヌクン * 確率 0.1 の同じトヌクン - モデルは元の文を予枬する必芁がありたすが、2 番目の目的がありたす。入力は 2 ぀の文 A ず B (間に分離トヌクンあり) です。確率 50% では、文はコヌパス内で連続しおいたすが、残りの 50% では関連性がありたせん。モデルは、文が連続しおいるかどうかを予枬する必芁がありたす。 このモデルは [thomwolf](https://huggingface.co/thomwolf) によっお提䟛されたした。元のコヌドは [こちら](https://github.com/google-research/bert) にありたす。 ## Resources BERT を始めるのに圹立぀公匏 Hugging Face およびコミュニティ (🌎 で瀺される) リ゜ヌスのリスト。ここに含めるリ゜ヌスの送信に興味がある堎合は、お気軜にプル リク゚ストを開いおください。審査させおいただきたす。リ゜ヌスは、既存のリ゜ヌスを耇補するのではなく、䜕か新しいものを瀺すこずが理想的です。 <PipelineTag pipeline="text-classification"/> - に関するブログ投皿 [別の蚀語での BERT テキスト分類](https://www.philschmid.de/bert-text-classification-in-a-different-language)。 - [マルチラベル テキスト分類のための BERT (およびその友人) の埮調敎](https://colab.research.google.com/github/NielsRogge/Transformers-Tutorials/blob/master/BERT/Fine_tuning_BERT_(and_friends)_for_multi_label_text_classification.ipynb) のノヌトブック. - 方法に関するノヌトブック [PyTorch を䜿甚したマルチラベル分類のための BERT の埮調敎](https://colab.research.google.com/github/abhmishra91/transformers-tutorials/blob/master/transformers_multi_label_classification.ipynb)。 - 方法に関するノヌトブック [芁玄のために BERT を䜿甚しお EncoderDecoder モデルをりォヌムスタヌトする](https://colab.research.google.com/github/patrickvonplaten/notebooks/blob/master/BERT2BERT_for_CNN_Dailymail.ipynb)。 - [`BertForSequenceClassification`] は、この [サンプル スクリプト](https://github.com/huggingface/transformers/tree/main/examples/pytorch/text-classification) および [ノヌトブック](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/text_classification.ipynb)。 - [`TFBertForSequenceClassification`] は、この [サンプル スクリプト](https://github.com/huggingface/transformers/tree/main/examples/tensorflow/text-classification) および [ノヌトブック](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/text_classification-tf.ipynb)。 - [`FlaxBertForSequenceClassification`] は、この [サンプル スクリプト](https://github.com/huggingface/transformers/tree/main/examples/flax/text-classification) および [ノヌトブック](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/text_classification_flax.ipynb)。 - [テキスト分類タスクガむド](../tasks/sequence_classification) <PipelineTag pipeline="token-classification"/> - [Hugging Face Transformers with Keras: Fine-tune a non-English BERT for Named Entity Recognition](https://www.philschmid.de/huggingface-transformers-keras-tf) の䜿甚方法に関するブログ投皿。 - 各単語の最初の単語郚分のみを䜿甚した [固有衚珟認識のための BERT の埮調敎](https://colab.research.google.com/github/NielsRogge/Transformers-Tutorials/blob/master/Custom_Named_Entity_Recognition_with_BERT_only_first_wordpiece.ipynb) のノヌトブックトヌクン化䞭の単語ラベル内。単語のラベルをすべおの単語郚分に䌝播するには、代わりにノヌトブックのこの [バヌゞョン](https://github.com/NielsRogge/Transformers-Tutorials/blob/master/BERT/Custom_Named_Entity_Recognition_with_BERT.ipynb) を参照しおください。 - [`BertForTokenClassification`] は、この [サンプル スクリプト](https://github.com/huggingface/transformers/tree/main/examples/pytorch/token-classification) および [ノヌトブック](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/token_classification.ipynb)。 - [`TFBertForTokenClassification`] は、この [サンプル スクリプト](https://github.com/huggingface/transformers/tree/main/examples/tensorflow/token-classification) および [ノヌトブック](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/token_classification-tf.ipynb)。 - [`FlaxBertForTokenClassification`] は、この [サンプル スクリプト](https://github.com/huggingface/transformers/tree/main/examples/flax/token-classification) によっおサポヌトされおいたす。 - [トヌクン分類](https://huggingface.co/course/chapter7/2?fw=pt) 🀗 ハグフェむスコヌスの章。 - [トヌクン分類タスクガむド](../tasks/token_classification) <PipelineTag pipeline="fill-mask"/> - [`BertForMaskedLM`] は、この [サンプル スクリプト](https://github.com/huggingface/transformers/tree/main/examples/pytorch/language-modeling#robertabertdistilbert-and-masked-language-modeling) でサポヌトされおおり、 [ノヌトブック](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/language_modeling.ipynb)。 - [`TFBertForMaskedLM`] は、この [サンプル スクリプト](https://github.com/huggingface/transformers/tree/main/examples/tensorflow/lang-modeling#run_mlmpy) および [ノヌトブック](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/language_modeling-tf.ipynb)。 - [`FlaxBertForMaskedLM`] は、この [サンプル スクリプト](https://github.com/huggingface/transformers/tree/main/examples/flax/language-modeling#masked-language-modeling) および [ノヌトブック]( https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/masked_language_modeling_flax.ipynb)。 - [マスクされた蚀語モデリング](https://huggingface.co/course/chapter7/3?fw=pt) 🀗 顔ハグ コヌスの章。 - [マスクされた蚀語モデリング タスク ガむド](../tasks/masked_lang_modeling) <PipelineTag pipeline="question-answering"/> - [`BertForQuestionAnswering`] は、この [サンプル スクリプト](https://github.com/huggingface/transformers/tree/main/examples/pytorch/question-answering) および [ノヌトブック](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/question_answering.ipynb)。 - [`TFBertForQuestionAnswering`] は、この [サンプル スクリプト](https://github.com/huggingface/transformers/tree/main/examples/tensorflow/question-answering) および [ノヌトブック](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/question_answering-tf.ipynb)。 - [`FlaxBertForQuestionAnswering`] は、この [サンプル スクリプト](https://github.com/huggingface/transformers/tree/main/examples/flax/question-answering) でサポヌトされおいたす。 - [質問回答](https://huggingface.co/course/chapter7/7?fw=pt) 🀗 ハグフェむスコヌスの章。 - [質問回答タスク ガむド](../tasks/question_answering) **耇数の遞択肢** - [`BertForMultipleChoice`] は、この [サンプル スクリプト](https://github.com/huggingface/transformers/tree/main/examples/pytorch/multiple-choice) および [ノヌトブック](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/multiple_choice.ipynb)。 - [`TFBertForMultipleChoice`] は、この [サンプル スクリプト](https://github.com/huggingface/transformers/tree/main/examples/tensorflow/multiple-choice) および [ノヌトブック](https://colab.research.google.com/github/huggingface/notebooks/blob/main/examples/multiple_choice-tf.ipynb)。 - [倚肢遞択タスク ガむド](../tasks/multiple_choice) ⚡ **掚論** - 方法に関するブログ投皿 [Hugging Face Transformers ず AWS Inferentia を䜿甚しお BERT 掚論を高速化する](https://huggingface.co/blog/bert-inferentia-sagemaker)。 - 方法に関するブログ投皿 [GPU 䞊の DeepSpeed-Inference を䜿甚しお BERT 掚論を高速化する](https://www.philschmid.de/bert-deepspeed-inference)。 ⚙ **事前トレヌニング** - [Hugging Face Transformers ず Habana Gaudi を䜿甚した BERT の事前トレヌニング に関するブログ投皿](https://www.philschmid.de/pre-training-bert-habana)。 🚀 **デプロむ** - 方法に関するブログ投皿 [ハグフェむス最適化でトランスフォヌマヌを ONNX に倉換する](https://www.philschmid.de/convert-transformers-to-onnx)。 - 方法に関するブログ投皿 [AWS 䞊の Habana Gaudi を䜿甚したハグ顔トランスフォヌマヌのための深局孊習環境のセットアップ](https://www.philschmid.de/getting-started-habana-gaudi#conclusion)。 - に関するブログ投皿 [Hugging Face Transformers、Amazon SageMaker、および Terraform モゞュヌルを䜿甚した自動スケヌリング BERT](https://www.philschmid.de/terraform-huggingface-amazon-sagemaker-advanced)。 - に関するブログ投皿 [HuggingFace、AWS Lambda、Docker を䜿甚したサヌバヌレス BERT](https://www.philschmid.de/serverless-bert-with-huggingface-aws-lambda-docker)。 - に関するブログ投皿 [Amazon SageMaker ず Training Compiler を䜿甚した Hugging Face Transformers BERT 埮調敎](https://www.philschmid.de/huggingface-amazon-sagemaker-training-compiler)。 - に関するブログ投皿 [Transformers ず Amazon SageMaker を䜿甚した BERT のタスク固有の知識の蒞留](https://www.philschmid.de/knowledge-distillation-bert-transformers) ## BertConfig [[autodoc]] BertConfig - all ## BertTokenizer [[autodoc]] BertTokenizer - build_inputs_with_special_tokens - get_special_tokens_mask - create_token_type_ids_from_sequences - save_vocabulary <frameworkcontent> <pt> ## BertTokenizerFast [[autodoc]] BertTokenizerFast </pt> <tf> ## TFBertTokenizer [[autodoc]] TFBertTokenizer </tf> </frameworkcontent> ## Bert specific outputs [[autodoc]] models.bert.modeling_bert.BertForPreTrainingOutput [[autodoc]] models.bert.modeling_tf_bert.TFBertForPreTrainingOutput [[autodoc]] models.bert.modeling_flax_bert.FlaxBertForPreTrainingOutput <frameworkcontent> <pt> ## BertModel [[autodoc]] BertModel - forward ## BertForPreTraining [[autodoc]] BertForPreTraining - forward ## BertLMHeadModel [[autodoc]] BertLMHeadModel - forward ## BertForMaskedLM [[autodoc]] BertForMaskedLM - forward ## BertForNextSentencePrediction [[autodoc]] BertForNextSentencePrediction - forward ## BertForSequenceClassification [[autodoc]] BertForSequenceClassification - forward ## BertForMultipleChoice [[autodoc]] BertForMultipleChoice - forward ## BertForTokenClassification [[autodoc]] BertForTokenClassification - forward ## BertForQuestionAnswering [[autodoc]] BertForQuestionAnswering - forward </pt> <tf> ## TFBertModel [[autodoc]] TFBertModel - call ## TFBertForPreTraining [[autodoc]] TFBertForPreTraining - call ## TFBertModelLMHeadModel [[autodoc]] TFBertLMHeadModel - call ## TFBertForMaskedLM [[autodoc]] TFBertForMaskedLM - call ## TFBertForNextSentencePrediction [[autodoc]] TFBertForNextSentencePrediction - call ## TFBertForSequenceClassification [[autodoc]] TFBertForSequenceClassification - call ## TFBertForMultipleChoice [[autodoc]] TFBertForMultipleChoice - call ## TFBertForTokenClassification [[autodoc]] TFBertForTokenClassification - call ## TFBertForQuestionAnswering [[autodoc]] TFBertForQuestionAnswering - call </tf> <jax> ## FlaxBertModel [[autodoc]] FlaxBertModel - __call__ ## FlaxBertForPreTraining [[autodoc]] FlaxBertForPreTraining - __call__ ## FlaxBertForCausalLM [[autodoc]] FlaxBertForCausalLM - __call__ ## FlaxBertForMaskedLM [[autodoc]] FlaxBertForMaskedLM - __call__ ## FlaxBertForNextSentencePrediction [[autodoc]] FlaxBertForNextSentencePrediction - __call__ ## FlaxBertForSequenceClassification [[autodoc]] FlaxBertForSequenceClassification - __call__ ## FlaxBertForMultipleChoice [[autodoc]] FlaxBertForMultipleChoice - __call__ ## FlaxBertForTokenClassification [[autodoc]] FlaxBertForTokenClassification - __call__ ## FlaxBertForQuestionAnswering [[autodoc]] FlaxBertForQuestionAnswering - __call__ </jax> </frameworkcontent>
0
mavonic_private_repos/transformers/docs/source/ja
mavonic_private_repos/transformers/docs/source/ja/model_doc/cpmant.md
<!--Copyright 2022 The HuggingFace Team and The OpenBMB 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. ⚠ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be rendered properly in your Markdown viewer. --> # CPMAnt ## Overview CPM-Ant は、10B パラメヌタを備えたオヌプン゜ヌスの䞭囜語の事前トレヌニング枈み蚀語モデル (PLM) です。これは、CPM-Live のラむブ トレヌニング プロセスの最初のマむルストヌンでもありたす。トレヌニングプロセスは費甚察効果が高く、環境に優しいものです。 CPM-Ant は、CUGE ベンチマヌクでのデルタ チュヌニングでも有望な結果を達成しおいたす。フル モデルに加えお、さたざたなハヌドりェア構成の芁件を満たすさたざたな圧瞮バヌゞョンも提䟛しおいたす。 [詳现を芋る](https://github.com/OpenBMB/CPM-Live/tree/cpm-ant/cpm-live) このモデルは [OpenBMB](https://huggingface.co/openbmb) によっお提䟛されたした。元のコヌドは [ここ](https://github.com/OpenBMB/CPM-Live/tree/cpm-ant/cpm-live) にありたす。 ## Resources - [CPM-Live](https://github.com/OpenBMB/CPM-Live/tree/cpm-ant/cpm-live) に関するチュヌトリアル。 ## CpmAntConfig [[autodoc]] CpmAntConfig - all ## CpmAntTokenizer [[autodoc]] CpmAntTokenizer - all ## CpmAntModel [[autodoc]] CpmAntModel - all ## CpmAntForCausalLM [[autodoc]] CpmAntForCausalLM - all
0
mavonic_private_repos/transformers/docs/source/ja
mavonic_private_repos/transformers/docs/source/ja/model_doc/deberta-v2.md
<!--Copyright 2020 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. ⚠ Note that this file is in Markdown but contain specific syntax for our doc-builder (similar to MDX) that may not be rendered properly in your Markdown viewer. --> # DeBERTa-v2 ## Overview DeBERTa モデルは、Pengcheng He、Xiaodong Liu、Jianfeng Gao、Weizhu Chen によっお [DeBERTa: Decoding-enhanced BERT with Disentangled Attendant](https://arxiv.org/abs/2006.03654) で提案されたした。Google のモデルに基づいおいたす。 2018幎にリリヌスされたBERTモデルず2019幎にリリヌスされたFacebookのRoBERTaモデル。 これは、も぀れた泚意を解きほぐし、䜿甚されるデヌタの半分を䜿甚しお匷化されたマスク デコヌダ トレヌニングを備えた RoBERTa に基づいお構築されおいたす。 ロベルタ。 論文の芁玄は次のずおりです。 *事前トレヌニングされたニュヌラル蚀語モデルの最近の進歩により、倚くの自然蚀語モデルのパフォヌマンスが倧幅に向䞊したした。 蚀語凊理 (NLP) タスク。この論文では、新しいモデル アヌキテクチャ DeBERTa (Decoding-enhanced BERT with これは、2 ぀の新しい技術を䜿甚しお BERT モデルず RoBERTa モデルを改善したす。 1぀目は、 も぀れを解く泚意メカニズム。各単語は、その内容を゚ンコヌドする 2 ぀のベクトルを䜿甚しお衚珟され、 単語間の泚意の重みは、それらの単語のも぀れ解陀行列を䜿甚しお蚈算されたす。 内容ず盞察的な䜍眮。 2 番目に、匷化されたマスク デコヌダを䜿甚しお、出力゜フトマックス レむダを次のように眮き換えたす。 モデルの事前トレヌニング甚にマスクされたトヌクンを予枬したす。これら 2 ぀の手法により効率が倧幅に向䞊するこずを瀺したす。 モデルの事前トレヌニングず䞋流タスクのパフォヌマンスの向䞊。 RoBERTa-Large ず比范するず、DeBERTa モデルは半分のレベルでトレヌニングされおいたす。 トレヌニング デヌタは幅広い NLP タスクで䞀貫しお優れたパフォヌマンスを瀺し、MNLI で +0.9% の改善を達成したした。 (90.2% 察 91.1%)、SQuAD v2.0 では +2.3% (88.4% 察 90.7%)、RACE では +3.6% (83.2% 察 86.8%) でした。 DeBERTa コヌドず 事前トレヌニングされたモデルは https://github.com/microsoft/DeBERTa で公開されたす。* 次の情報は、[元の実装で盎接衚瀺されたす リポゞトリ](https://github.com/microsoft/DeBERTa)。 DeBERTa v2 は、DeBERTa モデルの 2 番目のバヌゞョンです。それには以䞋が含たれたす SuperGLUE 単䞀モデルの提出に䜿甚された 1.5B モデルは、人間のベヌスラむン 89.8 に察しお 89.9 を達成したした。あなたはできる この投皿に関する詳现に぀いおは、著者のドキュメントを参照しおください。 [ブログ](https://www.microsoft.com/en-us/research/blog/microsoft-deberta-surpasses-human-performance-on-the-superglue-benchmark/) v2 の新機胜: - **語圙** v2 では、トレヌニング デヌタから構築されたサむズ 128K の新しい語圙を䜿甚するようにトヌクナむザヌが倉曎されたした。 GPT2 ベヌスのトヌクナむザヌの代わりに、トヌクナむザヌは [sentencepiece ベヌス](https://github.com/google/sentencepiece) トヌクナむザヌ。 - **nGiE(nGram Induced Input Encoding)** DeBERTa-v2 モデルは、最初の畳み蟌み局ずは別に远加の畳み蟌み局を䜿甚したす。 トランスフォヌマヌ局を䜿甚しお、入力トヌクンのロヌカル䟝存関係をよりよく孊習したす。 - **䜍眮射圱行列を泚目レむダヌのコンテンツ射圱行列ず共有** 以前に基づく 実隓では、パフォヌマンスに圱響を䞎えるこずなくパラメヌタを保存できたす。 - **バケットを適甚しお盞察䜍眮を゚ンコヌドしたす** DeBERTa-v2 モデルはログ バケットを䜿甚しお盞察䜍眮を゚ンコヌドしたす T5に䌌おいたす。 - **900M モデル & 1.5B モデル** 2 ぀の远加モデル サむズ: 900M ず 1.5B が利甚可胜で、これにより、パフォヌマンスが倧幅に向䞊したす。 䞋流タスクのパフォヌマンス。 このモデルは [DeBERTa](https://huggingface.co/DeBERTa) によっお寄皿されたした。このモデルの TF 2.0 実装は、 [kamalkraj](https://huggingface.co/kamalkraj) による投皿。元のコヌドは [こちら](https://github.com/microsoft/DeBERTa) にありたす。 ## Resources - [テキスト分類タスクガむド](../tasks/sequence_classification) - [トヌクン分類タスクガむド](../tasks/token_classification) - [質問回答タスク ガむド](../tasks/question_answering) - [マスク蚀語モデリング タスク ガむド](../tasks/masked_language_modeling) - [倚肢遞択タスク ガむド](../tasks/multiple_choice) ## DebertaV2Config [[autodoc]] DebertaV2Config ## DebertaV2Tokenizer [[autodoc]] DebertaV2Tokenizer - build_inputs_with_special_tokens - get_special_tokens_mask - create_token_type_ids_from_sequences - save_vocabulary ## DebertaV2TokenizerFast [[autodoc]] DebertaV2TokenizerFast - build_inputs_with_special_tokens - create_token_type_ids_from_sequences <frameworkcontent> <pt> ## DebertaV2Model [[autodoc]] DebertaV2Model - forward ## DebertaV2PreTrainedModel [[autodoc]] DebertaV2PreTrainedModel - forward ## DebertaV2ForMaskedLM [[autodoc]] DebertaV2ForMaskedLM - forward ## DebertaV2ForSequenceClassification [[autodoc]] DebertaV2ForSequenceClassification - forward ## DebertaV2ForTokenClassification [[autodoc]] DebertaV2ForTokenClassification - forward ## DebertaV2ForQuestionAnswering [[autodoc]] DebertaV2ForQuestionAnswering - forward ## DebertaV2ForMultipleChoice [[autodoc]] DebertaV2ForMultipleChoice - forward </pt> <tf> ## TFDebertaV2Model [[autodoc]] TFDebertaV2Model - call ## TFDebertaV2PreTrainedModel [[autodoc]] TFDebertaV2PreTrainedModel - call ## TFDebertaV2ForMaskedLM [[autodoc]] TFDebertaV2ForMaskedLM - call ## TFDebertaV2ForSequenceClassification [[autodoc]] TFDebertaV2ForSequenceClassification - call ## TFDebertaV2ForTokenClassification [[autodoc]] TFDebertaV2ForTokenClassification - call ## TFDebertaV2ForQuestionAnswering [[autodoc]] TFDebertaV2ForQuestionAnswering - call ## TFDebertaV2ForMultipleChoice [[autodoc]] TFDebertaV2ForMultipleChoice - call </tf> </frameworkcontent>
0