mirror of
https://github.com/huggingface/transformers.git
synced 2025-07-05 05:40:05 +06:00
9b7668c03a
* udpaet * update * Update docs/source/ja/autoclass_tutorial.md Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com> * add codes workflows/build_pr_documentation.yml * Create preprocessing.md * added traning.md * Create Model_sharing.md * add quicktour.md * new * ll * Create benchmark.md * Create Tensorflow_model * add * add community.md * add create_a_model * create custom_model.md * create_custom_tools.md * create fast_tokenizers.md * create * add * Update docs/source/ja/_toctree.yml Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com> * md * add * commit * add * h * Update docs/source/ja/peft.md Co-authored-by: Younes Belkada <49240599+younesbelkada@users.noreply.github.com> * Update docs/source/ja/_toctree.yml Co-authored-by: Younes Belkada <49240599+younesbelkada@users.noreply.github.com> * Update docs/source/ja/_toctree.yml Co-authored-by: Younes Belkada <49240599+younesbelkada@users.noreply.github.com> * Suggested Update * add perf_train_gpu_one.md * added perf based MD files * Modify toctree.yml and Add transmartion to md codes * Add `serialization.md` and edit `_toctree.yml` * add task summary and tasks explained * Add and Modify files starting from T * Add testing.md * Create main_classes files * delete main_classes folder * Add toctree.yml * Update llm_tutorail.md * Update docs/source/ja/_toctree.yml Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com> * Update misspelled filenames * Update docs/source/ja/_toctree.yml Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com> * Update docs/source/ja/_toctree.yml * Update docs/source/ja/_toctree.yml * missplled file names inmrpovements * Update _toctree.yml * close tip block * close another tip block * Update docs/source/ja/quicktour.md Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com> * Update docs/source/ja/pipeline_tutorial.md Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com> * Update docs/source/ja/pipeline_tutorial.md Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com> * Update docs/source/ja/preprocessing.md Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com> * Update docs/source/ja/peft.md Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com> * Update docs/source/ja/add_new_model.md Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com> * Update docs/source/ja/testing.md Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com> * Update docs/source/ja/task_summary.md Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com> * Update docs/source/ja/tasks_explained.md Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com> * Update glossary.md * Update docs/source/ja/transformers_agents.md Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com> * Update docs/source/ja/llm_tutorial.md Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com> * Update docs/source/ja/create_a_model.md Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com> * Update docs/source/ja/torchscript.md Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com> * Update docs/source/ja/benchmarks.md Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com> * Update docs/source/ja/troubleshooting.md Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com> * Update docs/source/ja/troubleshooting.md Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com> * Update docs/source/ja/troubleshooting.md Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com> * Update docs/source/ja/add_new_model.md Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com> * Update perf_torch_compile.md * Update Year to default in en documentation * Final Update --------- Co-authored-by: Steven Liu <59462357+stevhliu@users.noreply.github.com> Co-authored-by: Younes Belkada <49240599+younesbelkada@users.noreply.github.com>
35 lines
2.8 KiB
Markdown
35 lines
2.8 KiB
Markdown
<!--Copyright 2023 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.
|
||
|
||
-->
|
||
|
||
|
||
# BERTology
|
||
|
||
大規模なトランスフォーマー、例えばBERTの内部動作を調査する研究領域が急成長しています（これを「BERTology」とも呼びます）。この分野の良い例は以下です：
|
||
|
||
- BERT Rediscovers the Classical NLP Pipeline by Ian Tenney, Dipanjan Das, Ellie Pavlick:
|
||
[論文リンク](https://arxiv.org/abs/1905.05950)
|
||
- Are Sixteen Heads Really Better than One? by Paul Michel, Omer Levy, Graham Neubig: [論文リンク](https://arxiv.org/abs/1905.10650)
|
||
- What Does BERT Look At? An Analysis of BERT's Attention by Kevin Clark, Urvashi Khandelwal, Omer Levy, Christopher D. Manning: [論文リンク](https://arxiv.org/abs/1906.04341)
|
||
- CAT-probing: A Metric-based Approach to Interpret How Pre-trained Models for Programming Language Attend Code Structure: [論文リンク](https://arxiv.org/abs/2210.04633)
|
||
|
||
この新しい分野の発展を支援するために、BERT/GPT/GPT-2モデルにいくつかの追加機能を組み込み、人々が内部表現にアクセスできるようにしました。これらの機能は、主にPaul Michel氏の優れた研究（[論文リンク](https://arxiv.org/abs/1905.10650)）に基づいています。具体的には、以下の機能が含まれています：
|
||
|
||
- BERT/GPT/GPT-2のすべての隠れ状態にアクセスすることができます。
|
||
- BERT/GPT/GPT-2の各ヘッドの注意重みにアクセスできます。
|
||
- ヘッドの出力値と勾配を取得し、ヘッドの重要性スコアを計算し、[論文リンク](https://arxiv.org/abs/1905.10650)で説明されているようにヘッドを削減できます。
|
||
|
||
これらの機能を理解し、使用するのを支援するために、特定のサンプルスクリプト「[bertology.py](https://github.com/huggingface/transformers/tree/main/examples/research_projects/bertology/run_bertology.py)」を追加しました。このスクリプトは、GLUEで事前トレーニングされたモデルから情報を抽出し、ヘッドを削減する役割を果たします。
|