Remove more unused attributes in config classes (#21543)

* Remove unused decoder_layerdrop

* Update SPECIAL_CASES_TO_ALLOW for MT5Config

* Remove unused position_embedding_init_scale

* Remove unused decoder_max_relative_position

* Use unused decoder_max_relative_position

* Remove unused init_std

* Remove unused forgotten attributes

* Remove unused patch_norm

* Remove unused max_seq_len

* Update SPECIAL_CASES_TO_ALLOW for OneFormerConfig

---------

Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>
This commit is contained in:
Yih-Dar
2023-02-10 22:57:28 +01:00
committed by GitHub
parent 862e8e4f4a
commit b47a16743b
15 changed files with 7 additions and 54 deletions

View File

@@ -45,13 +45,17 @@ SPECIAL_CASES_TO_ALLOW = {
"EsmConfig": ["is_folding_model"],
# used during training (despite we don't have training script for these models yet)
"Mask2FormerConfig": ["ignore_value"],
# used during training (despite we don't have training script for these models yet)
"OneFormerConfig": ["ignore_value"],
# `ignore_value` used during training (despite we don't have training script for these models yet)
# `norm` used in conversion script (despite not using in the modeling file)
"OneFormerConfig": ["ignore_value", "norm"],
# used during preprocessing and collation, see `collating_graphormer.py`
"GraphormerConfig": ["spatial_pos_max"],
# used internally in the configuration class file
"T5Config": ["feed_forward_proj"],
# used internally in the configuration class file
# `tokenizer_class` get default value `T5Tokenizer` intentionally
"MT5Config": ["feed_forward_proj", "tokenizer_class"],
# used internally in the configuration class file
"LongT5Config": ["feed_forward_proj"],
# used internally in the configuration class file
"SwitchTransformersConfig": ["feed_forward_proj"],