From 28ed53db6fa2582f3d9c73c3a507c92d73f5917d Mon Sep 17 00:00:00 2001 From: Mu Yang Date: Wed, 6 Jan 2021 13:53:05 +0800 Subject: [PATCH 1/6] Add #parameters to readme. --- README.rst | 34 +++++++++++++++++----------------- 1 file changed, 17 insertions(+), 17 deletions(-) diff --git a/README.rst b/README.rst index 2e7db11..0d8d16c 100644 --- a/README.rst +++ b/README.rst @@ -167,20 +167,20 @@ Model Performance | 以下是我們的模型與其他的模型之性能比較。 | 各個任務皆測試於繁體中文的測試集。 -================================ =========== ======== ========== ========= -Model Perplexity† WS (F1)‡ POS (ACC)‡ NER (F1)‡ -================================ =========== ======== ========== ========= -ckiplab/albert-tiny-chinese 4.80 96.66% 94.48% 71.17% -ckiplab/albert-base-chinese 2.65 97.33% 95.30% 79.47% -ckiplab/bert-base-chinese 1.88 97.60% 95.67% 81.18% -ckiplab/gpt2-base-chinese 14.40 -- -- -- --------------------------------- ----------- -------- ---------- --------- - --------------------------------- ----------- -------- ---------- --------- -voidful/albert_chinese_tiny 74.93 -- -- -- -voidful/albert_chinese_base 22.34 -- -- -- -bert-base-chinese 2.53 -- -- -- -================================ =========== ======== ========== ========= +================================ =========== =========== ======== ========== ========= +Model #Parameters Perplexity† WS (F1)‡ POS (ACC)‡ NER (F1)‡ +================================ =========== =========== ======== ========== ========= +ckiplab/albert-tiny-chinese 4,044,312 4.80 96.66% 94.48% 71.17% +ckiplab/albert-base-chinese 10,077,192 2.65 97.33% 95.30% 79.47% +ckiplab/bert-base-chinese 102,290,312 1.88 97.60% 95.67% 81.18% +ckiplab/gpt2-base-chinese 102,068,736 14.40 -- -- -- +-------------------------------- ----------- ----------- -------- ---------- --------- + +-------------------------------- ----------- ----------- -------- ---------- --------- +voidful/albert_chinese_tiny 4,080,520 74.93 -- -- -- +voidful/albert_chinese_base 10,547,968 22.34 -- -- -- +bert-base-chinese 102,267,648 2.53 -- -- -- +================================ =========== =========== ======== ========== ========= | † Perplexity; the smaller the better. | † 混淆度；數字越小越好。 @@ -197,10 +197,10 @@ Training Corpus | Chinese Wikipedia text (20200801 dump), translated to Traditional using `OpenCC `_. | 中文維基的文章（20200801 版本），利用 `OpenCC `_ 翻譯成繁體中文。 * CNA: https://catalog.ldc.upenn.edu/LDC2011T13 - | Chinese Gigaword Fifth Edition — CNA (Central News Agency part). + | Chinese Gigaword Fifth Edition — CNA (Central News Agency) part. | 中文 Gigaword 第五版 — CNA（中央社）的部分. * ASBC: http://asbc.iis.sinica.edu.tw - | Academia Sinica Balanced Corpus of Modern Chinese version 4. + | Academia Sinica Balanced Corpus of Modern Chinese release 4.0. | 中央研究院漢語平衡語料庫第四版。 * OntoNotes: https://catalog.ldc.upenn.edu/LDC2013T19 | OntoNotes release 5.0, Chinese part, translated to Traditional using `OpenCC `_. @@ -292,7 +292,7 @@ NLP Tools Usage 2. Load models """""""""""""" -| We provide three levels (1–3) of drivers. Level 1 if the fastest, and level 3 (default) is the most accurate. +| We provide three levels (1–3) of drivers. Level 1 is the fastest, and level 3 (default) is the most accurate. | 我們的工具分為三個等級（1—3）。等級一最快，等級三（預設值）最精準。 .. code-block:: python From f4a9039e9d2c1ff63173dfa928c0f09072761bb0 Mon Sep 17 00:00:00 2001 From: Mu Yang Date: Wed, 6 Jan 2021 14:06:57 +0800 Subject: [PATCH 2/6] Update documentation. --- README.rst | 9 ++------- 1 file changed, 2 insertions(+), 7 deletions(-) diff --git a/README.rst b/README.rst index 0d8d16c..f39fcdc 100644 --- a/README.rst +++ b/README.rst @@ -123,17 +123,12 @@ Model Usage from transformers import ( BertTokenizerFast, - AutoModelForMaskedLM, + AutoModel, AutoModelForTokenClassification, ) - # language model tokenizer = BertTokenizerFast.from_pretrained('bert-base-chinese') - model = AutoModelForMaskedLM.from_pretrained('ckiplab/albert-tiny-chinese') # or other models above - - # nlp task model - tokenizer = BertTokenizerFast.from_pretrained('bert-base-chinese') - model = AutoModelForTokenClassification.from_pretrained('ckiplab/albert-tiny-chinese-ws') # or other models above + model = AutoModel.from_pretrained('ckiplab/albert-tiny-chinese') # or other models above Model Fine-Tunning ^^^^^^^^^^^^^^^^^^ From dcbe441682cab2247c86a5bde4aa15a9af705898 Mon Sep 17 00:00:00 2001 From: Mu Yang Date: Wed, 6 Jan 2021 16:48:32 +0800 Subject: [PATCH 3/6] Abbrivate parameters in documentation. --- README.rst | 14 +++++++------- 1 file changed, 7 insertions(+), 7 deletions(-) diff --git a/README.rst b/README.rst index f39fcdc..9a1ce86 100644 --- a/README.rst +++ b/README.rst @@ -165,16 +165,16 @@ Model Performance ================================ =========== =========== ======== ========== ========= Model #Parameters Perplexity† WS (F1)‡ POS (ACC)‡ NER (F1)‡ ================================ =========== =========== ======== ========== ========= -ckiplab/albert-tiny-chinese 4,044,312 4.80 96.66% 94.48% 71.17% -ckiplab/albert-base-chinese 10,077,192 2.65 97.33% 95.30% 79.47% -ckiplab/bert-base-chinese 102,290,312 1.88 97.60% 95.67% 81.18% -ckiplab/gpt2-base-chinese 102,068,736 14.40 -- -- -- +ckiplab/albert-tiny-chinese 4M 4.80 96.66% 94.48% 71.17% +ckiplab/albert-base-chinese 10M 2.65 97.33% 95.30% 79.47% +ckiplab/bert-base-chinese 102M 1.88 97.60% 95.67% 81.18% +ckiplab/gpt2-base-chinese 102M 14.40 -- -- -- -------------------------------- ----------- ----------- -------- ---------- --------- -------------------------------- ----------- ----------- -------- ---------- --------- -voidful/albert_chinese_tiny 4,080,520 74.93 -- -- -- -voidful/albert_chinese_base 10,547,968 22.34 -- -- -- -bert-base-chinese 102,267,648 2.53 -- -- -- +voidful/albert_chinese_tiny 4M 74.93 -- -- -- +voidful/albert_chinese_base 10M 22.34 -- -- -- +bert-base-chinese 102M 2.53 -- -- -- ================================ =========== =========== ======== ========== ========= | † Perplexity; the smaller the better. From 2a840bce98c131454aee79ecbe75532eea5e7ebe Mon Sep 17 00:00:00 2001 From: Mu Yang Date: Wed, 6 Jan 2021 17:08:34 +0800 Subject: [PATCH 4/6] Update usage in documentation. --- README.rst | 14 ++++++++++++-- 1 file changed, 12 insertions(+), 2 deletions(-) diff --git a/README.rst b/README.rst index 9a1ce86..aa13f69 100644 --- a/README.rst +++ b/README.rst @@ -123,12 +123,22 @@ Model Usage from transformers import ( BertTokenizerFast, - AutoModel, + AutoModelForMaskedLM, + AutoModelForCausalLM, AutoModelForTokenClassification, ) + # masked language model (ALBERT, BERT) tokenizer = BertTokenizerFast.from_pretrained('bert-base-chinese') - model = AutoModel.from_pretrained('ckiplab/albert-tiny-chinese') # or other models above + model = AutoModelForMaskedLM.from_pretrained('ckiplab/albert-tiny-chinese') # or other models above + + # casual language model (GPT2) + tokenizer = BertTokenizerFast.from_pretrained('bert-base-chinese') + model = AutoModelForCausalLM.from_pretrained('ckiplab/gpt2-base-chinese') # or other models above + + # nlp task model + tokenizer = BertTokenizerFast.from_pretrained('bert-base-chinese') + model = AutoModelForTokenClassification.from_pretrained('ckiplab/albert-tiny-chinese-ws') # or other models above Model Fine-Tunning ^^^^^^^^^^^^^^^^^^ From 970dba83206544a852b30edb96f3525280ac5ab4 Mon Sep 17 00:00:00 2001 From: Mu Yang Date: Mon, 1 Feb 2021 14:40:41 +0800 Subject: [PATCH 5/6] Update PyTorch version requirement. --- README.rst | 2 +- requirements.txt | 2 +- setup.py | 2 +- 3 files changed, 3 insertions(+), 3 deletions(-) diff --git a/README.rst b/README.rst index aa13f69..101a573 100644 --- a/README.rst +++ b/README.rst @@ -275,7 +275,7 @@ Installation Requirements: * `Python `__ 3.6+ -* `PyTorch `__ 1.1+ +* `PyTorch `__ 1.5+ * `HuggingFace Transformers `__ 3.5+ NLP Tools Usage diff --git a/requirements.txt b/requirements.txt index 150d077..52e51c2 100644 --- a/requirements.txt +++ b/requirements.txt @@ -1,3 +1,3 @@ -torch>=1.1.0 +torch>=1.5.0 tqdm>=4.27 transformers>=3.5.0 diff --git a/setup.py b/setup.py index 5a14130..a604cce 100644 --- a/setup.py +++ b/setup.py @@ -43,7 +43,7 @@ def main(): python_requires='>=3.6', packages=find_namespace_packages(include=['ckip_transformers', 'ckip_transformers.*',]), install_requires=[ - 'torch>=1.1.0', + 'torch>=1.5.0', 'tqdm>=4.27', 'transformers>=3.5.0', ], From fc69fa4ba8800fb1ce628bac6ff20740950072df Mon Sep 17 00:00:00 2001 From: Mu Yang Date: Mon, 1 Feb 2021 14:47:40 +0800 Subject: [PATCH 6/6] Change version number. --- ckip_transformers/__init__.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/ckip_transformers/__init__.py b/ckip_transformers/__init__.py index 15890ba..7ef2ffb 100644 --- a/ckip_transformers/__init__.py +++ b/ckip_transformers/__init__.py @@ -10,7 +10,7 @@ __copyright__ = '2020 CKIP Lab' __title__ = 'CKIP Transformers' -__version__ = '0.2.1' +__version__ = '0.2.2' __description__ = 'CKIP Transformers' __license__ = 'GPL-3.0'