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

Add "Leveraging Pretrained Checkpoints for Generation" Seq2Seq models. (#6594)

Browse Source 
* add conversion script

* improve conversion script

* make style

* add tryout files

* fix

* update

* add causal bert

* better names

* add tokenizer file as well

* finish causal_bert

* fix small bugs

* improve generate

* change naming

* renaming

* renaming

* renaming

* remove leftover files

* clean files

* add fix tokenizer

* finalize

* correct slow test

* update docs

* small fixes

* fix link

* adapt check repo

* apply sams and sylvains recommendations

* fix import

* implement Lysandres recommendations

* fix logger warn
This commit is contained in:
Patrick von Platen
2020-09-10 16:40:51 +02:00
committed by GitHub
parent d1691d90e5
commit 7fd1febf38
20 changed files with 1508 additions and 9 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
src/transformers/file_utils.py
View File

@@ -530,7 +530,7 @@ def add_code_sample_docstrings(*docstr, tokenizer_class=None, checkpoint=None, o
code_sample = TF_MASKED_LM_SAMPLE if is_tf_class else PT_MASKED_LM_SAMPLE
elif "LMHead" in model_class:
code_sample = TF_CAUSAL_LM_SAMPLE if is_tf_class else PT_CAUSAL_LM_SAMPLE
elif "Model" in model_class:
elif "Model" in model_class or "Encoder" in model_class:
code_sample = TF_BASE_MODEL_SAMPLE if is_tf_class else PT_BASE_MODEL_SAMPLE
else:
raise ValueError(f"Docstring can't be built for model {model_class}")