Remove unused `max_position_embeddings ` in config classes (#20836)

Removed unused max_position_embeddings in config classes
Co-authored-by: ydshieh <ydshieh@users.noreply.github.com>

src/transformers/models/conditional_detr/configuration_conditional_detr.py

@@ -138,7 +138,6 @@ class ConditionalDetrConfig(PretrainedConfig):
         self,
         num_channels=3,
         num_queries=300,
-        max_position_embeddings=1024,
         encoder_layers=6,
         encoder_ffn_dim=2048,
         encoder_attention_heads=8,
@@ -175,7 +174,6 @@ class ConditionalDetrConfig(PretrainedConfig):
     ):
         self.num_channels = num_channels
         self.num_queries = num_queries
-        self.max_position_embeddings = max_position_embeddings
         self.d_model = d_model
         self.encoder_ffn_dim = encoder_ffn_dim
         self.encoder_layers = encoder_layers

src/transformers/models/detr/configuration_detr.py

@@ -135,7 +135,6 @@ class DetrConfig(PretrainedConfig):
         self,
         num_channels=3,
         num_queries=100,
-        max_position_embeddings=1024,
         encoder_layers=6,
         encoder_ffn_dim=2048,
         encoder_attention_heads=8,
@@ -171,7 +170,6 @@ class DetrConfig(PretrainedConfig):
     ):
         self.num_channels = num_channels
         self.num_queries = num_queries
-        self.max_position_embeddings = max_position_embeddings
         self.d_model = d_model
         self.encoder_ffn_dim = encoder_ffn_dim
         self.encoder_layers = encoder_layers

src/transformers/models/funnel/configuration_funnel.py

@@ -75,9 +75,6 @@ class FunnelConfig(PretrainedConfig):
             The dropout probability for the attention probabilities.
         activation_dropout (`float`, *optional*, defaults to 0.0):
             The dropout probability used between the two layers of the feed-forward blocks.
-        max_position_embeddings (`int`, *optional*, defaults to 512):
-            The maximum sequence length that this model might ever be used with. Typically set this to something large
-            just in case (e.g., 512 or 1024 or 2048).
         type_vocab_size (`int`, *optional*, defaults to 3):
             The vocabulary size of the `token_type_ids` passed when calling [`FunnelModel`] or [`TFFunnelModel`].
         initializer_range (`float`, *optional*, defaults to 0.1):
@@ -121,7 +118,6 @@ class FunnelConfig(PretrainedConfig):
         hidden_dropout=0.1,
         attention_dropout=0.1,
         activation_dropout=0.0,
-        max_position_embeddings=512,
         type_vocab_size=3,
         initializer_range=0.1,
         initializer_std=None,
@@ -148,7 +144,6 @@ class FunnelConfig(PretrainedConfig):
         self.hidden_dropout = hidden_dropout
         self.attention_dropout = attention_dropout
         self.activation_dropout = activation_dropout
-        self.max_position_embeddings = max_position_embeddings
         self.type_vocab_size = type_vocab_size
         self.initializer_range = initializer_range
         self.initializer_std = initializer_std

src/transformers/models/table_transformer/configuration_table_transformer.py

@@ -137,7 +137,6 @@ class TableTransformerConfig(PretrainedConfig):
         self,
         num_channels=3,
         num_queries=100,
-        max_position_embeddings=1024,
         encoder_layers=6,
         encoder_ffn_dim=2048,
         encoder_attention_heads=8,
@@ -173,7 +172,6 @@ class TableTransformerConfig(PretrainedConfig):
     ):
         self.num_channels = num_channels
         self.num_queries = num_queries
-        self.max_position_embeddings = max_position_embeddings
         self.d_model = d_model
         self.encoder_ffn_dim = encoder_ffn_dim
         self.encoder_layers = encoder_layers