Cleanup some config attributes (#20554)

* Remove is_encoder_decoder from some vision models

* cleanup more

* cleanup more
Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>

src/transformers/models/beit/configuration_beit.py

@@ -129,7 +129,6 @@ class BeitConfig(PretrainedConfig):
         attention_probs_dropout_prob=0.0,
         initializer_range=0.02,
         layer_norm_eps=1e-12,
-        is_encoder_decoder=False,
         image_size=224,
         patch_size=16,
         num_channels=3,

src/transformers/models/big_bird/configuration_big_bird.py

@@ -119,7 +119,6 @@ class BigBirdConfig(PretrainedConfig):
         initializer_range=0.02,
         layer_norm_eps=1e-12,
         use_cache=True,
-        is_encoder_decoder=False,
         pad_token_id=0,
         bos_token_id=1,
         eos_token_id=2,
@@ -153,7 +152,6 @@ class BigBirdConfig(PretrainedConfig):
         self.type_vocab_size = type_vocab_size
         self.layer_norm_eps = layer_norm_eps
         self.use_cache = use_cache
-        self.is_encoder_decoder = is_encoder_decoder
 
         self.rescale_embeddings = rescale_embeddings
         self.attention_type = attention_type

src/transformers/models/canine/configuration_canine.py

@@ -105,7 +105,6 @@ class CanineConfig(PretrainedConfig):
         initializer_range=0.02,
         layer_norm_eps=1e-12,
         use_cache=True,
-        is_encoder_decoder=False,
         pad_token_id=0,
         bos_token_id=0xE000,
         eos_token_id=0xE001,

src/transformers/models/convbert/configuration_convbert.py

@@ -102,7 +102,6 @@ class ConvBertConfig(PretrainedConfig):
         self,
         vocab_size=30522,
         hidden_size=768,
-        is_encoder_decoder=False,
         num_hidden_layers=12,
         num_attention_heads=12,
         intermediate_size=3072,
@@ -125,7 +124,6 @@ class ConvBertConfig(PretrainedConfig):
     ):
         super().__init__(
             pad_token_id=pad_token_id,
-            is_encoder_decoder=is_encoder_decoder,
             bos_token_id=bos_token_id,
             eos_token_id=eos_token_id,
             **kwargs,

src/transformers/models/convnext/configuration_convnext.py

@@ -90,7 +90,6 @@ class ConvNextConfig(PretrainedConfig):
         hidden_act="gelu",
         initializer_range=0.02,
         layer_norm_eps=1e-12,
-        is_encoder_decoder=False,
         layer_scale_init_value=1e-6,
         drop_path_rate=0.0,
         image_size=224,

src/transformers/models/data2vec/configuration_data2vec_vision.py

@@ -128,7 +128,6 @@ class Data2VecVisionConfig(PretrainedConfig):
         attention_probs_dropout_prob=0.0,
         initializer_range=0.02,
         layer_norm_eps=1e-12,
-        is_encoder_decoder=False,
         image_size=224,
         patch_size=16,
         num_channels=3,

src/transformers/models/deit/configuration_deit.py

@@ -104,7 +104,6 @@ class DeiTConfig(PretrainedConfig):
         attention_probs_dropout_prob=0.0,
         initializer_range=0.02,
         layer_norm_eps=1e-12,
-        is_encoder_decoder=False,
         image_size=224,
         patch_size=16,
         num_channels=3,

src/transformers/models/glpn/configuration_glpn.py

@@ -109,7 +109,6 @@ class GLPNConfig(PretrainedConfig):
         initializer_range=0.02,
         drop_path_rate=0.1,
         layer_norm_eps=1e-6,
-        is_encoder_decoder=False,
         decoder_hidden_size=64,
         max_depth=10,
         head_in_index=-1,

src/transformers/models/perceiver/configuration_perceiver.py

@@ -136,7 +136,6 @@ class PerceiverConfig(PretrainedConfig):
         position_embedding_init_scale=0.02,
         initializer_range=0.02,
         layer_norm_eps=1e-12,
-        is_encoder_decoder=False,
         use_query_residual=True,
         vocab_size=262,
         max_position_embeddings=2048,

src/transformers/models/qdqbert/configuration_qdqbert.py

@@ -100,7 +100,6 @@ class QDQBertConfig(PretrainedConfig):
         initializer_range=0.02,
         layer_norm_eps=1e-12,
         use_cache=True,
-        is_encoder_decoder=False,
         pad_token_id=1,
         bos_token_id=0,
         eos_token_id=2,

src/transformers/models/rembert/configuration_rembert.py

@@ -111,7 +111,6 @@ class RemBertConfig(PretrainedConfig):
         initializer_range=0.02,
         layer_norm_eps=1e-12,
         use_cache=True,
-        is_encoder_decoder=False,
         pad_token_id=0,
         bos_token_id=312,
         eos_token_id=313,

src/transformers/models/segformer/configuration_segformer.py

@@ -121,7 +121,6 @@ class SegformerConfig(PretrainedConfig):
         drop_path_rate=0.1,
         layer_norm_eps=1e-6,
         decoder_hidden_size=256,
-        is_encoder_decoder=False,
         semantic_loss_ignore_index=255,
         **kwargs
     ):

src/transformers/models/trajectory_transformer/configuration_trajectory_transformer.py

@@ -136,7 +136,6 @@ class TrajectoryTransformerConfig(PretrainedConfig):
         layer_norm_eps=1e-12,
         kaiming_initializer_range=1,
         use_cache=True,
-        is_encoder_decoder=False,
         pad_token_id=1,
         bos_token_id=50256,
         eos_token_id=50256,

src/transformers/models/vilt/configuration_vilt.py

@@ -113,7 +113,6 @@ class ViltConfig(PretrainedConfig):
         attention_probs_dropout_prob=0.0,
         initializer_range=0.02,
         layer_norm_eps=1e-12,
-        is_encoder_decoder=False,
         image_size=384,
         patch_size=32,
         num_channels=3,

src/transformers/models/vit/configuration_vit.py

@@ -101,7 +101,6 @@ class ViTConfig(PretrainedConfig):
         attention_probs_dropout_prob=0.0,
         initializer_range=0.02,
         layer_norm_eps=1e-12,
-        is_encoder_decoder=False,
         image_size=224,
         patch_size=16,
         num_channels=3,

src/transformers/models/vit_mae/configuration_vit_mae.py

@@ -106,7 +106,6 @@ class ViTMAEConfig(PretrainedConfig):
         attention_probs_dropout_prob=0.0,
         initializer_range=0.02,
         layer_norm_eps=1e-12,
-        is_encoder_decoder=False,
         image_size=224,
         patch_size=16,
         num_channels=3,

templates/adding_a_new_model/cookiecutter-template-{{cookiecutter.modelname}}/configuration_{{cookiecutter.lowercase_modelname}}.py

@@ -163,7 +163,6 @@ class {{cookiecutter.camelcase_modelname}}Config(PretrainedConfig):
         initializer_range=0.02,
         layer_norm_eps=1e-12,
         use_cache=True,
-        is_encoder_decoder=False,
         {% else -%}
         vocab_size=50265,
         max_position_embeddings=1024,