Update configuration_footballbert.py

configuration_footballbert.py

@@ -1,80 +1,80 @@
-"""FootballBERT Configuration"""
-
-from transformers import PretrainedConfig
-
-class FootballBERTConfig(PretrainedConfig):
-    """
-    Configuration class for FootballBERT models.
-    
-    FootballBERT treats players as tokens and matches as sentences, learning
-    contextual embeddings for football players through masked player prediction.
-    
-    Args:
-        vocab_size (`int`, *optional*, defaults to 99944):
-            Vocabulary size of the FootballBERT model (number of unique players).
-        hidden_size (`int`, *optional*, defaults to 256):
-            Dimensionality of the encoder layers and the pooler layer.
-        num_hidden_layers (`int`, *optional*, defaults to 6):
-            Number of hidden layers in the Transformer encoder.
-        num_attention_heads (`int`, *optional*, defaults to 1):
-            Number of attention heads for each attention layer in the Transformer encoder.
-        intermediate_size (`int`, *optional*, defaults to 1024):
-            Dimensionality of the "intermediate" (often named feed-forward) layer in the Transformer encoder.
-        hidden_dropout_prob (`float`, *optional*, defaults to 0.05):
-            The dropout probability for all fully connected layers in the embeddings, encoder, and pooler.
-        max_position_embeddings (`int`, *optional*, defaults to 68):
-            The maximum sequence length (2 teams * 34 players max).
-        num_positions (`int`, *optional*, defaults to 1419):
-            Number of unique position labels.
-        num_seasons (`int`, *optional*, defaults to 23):
-            Number of unique seasons in the dataset.
-        pad_token_id (`int`, *optional*, defaults to 99944):
-            The id of the padding token.
-        mask_token_id (`int`, *optional*, defaults to 99943):
-            The id of the mask token.
-        position_pad_token_id (`int`, *optional*, defaults to 1419):
-            The id of the padding token for positions.
-        season_pad_token_id (`int`, *optional*, defaults to 23):
-            The id of the padding token for seasons.
-    """
-    
-    model_type = "footballbert"
-    
-    def __init__(
-        self,
-        vocab_size: int = 99944,
-        hidden_size: int = 256,
-        num_hidden_layers: int = 6,
-        num_attention_heads: int = 1,
-        intermediate_size: int = 1024,
-        hidden_dropout_prob: float = 0.05,
-        max_position_embeddings: int = 68,
-        num_positions: int = 1419,
-        num_seasons: int = 23,
-        pad_token_id: int = 99944,
-        mask_token_id: int = 99943,
-        position_pad_token_id: int = 1419,
-        season_pad_token_id: int = 23,
-        **kwargs
-    ):
-        super().__init__(pad_token_id=pad_token_id, **kwargs)
-        
-        self.vocab_size = vocab_size
-        self.hidden_size = hidden_size
-        self.num_hidden_layers = num_hidden_layers
-        self.num_attention_heads = num_attention_heads
-        self.intermediate_size = intermediate_size
-        self.hidden_dropout_prob = hidden_dropout_prob
-        self.max_position_embeddings = max_position_embeddings
-        self.num_positions = num_positions
-        self.num_seasons = num_seasons
-        self.mask_token_id = mask_token_id
-        self.position_pad_token_id = position_pad_token_id
-        self.season_pad_token_id = season_pad_token_id
-
-        # auto mapping for model
-        self.auto_map = {
-            "AutoConfig": "configuration_footballbert.FootballBERTConfig",
-            "AutoModel": "modeling_footballbert.FootballBERTModel",
-            "AutoModelForMaskedLM": "modeling_footballbert.FootballBERTForMaskedPlayerPrediction",
+"""FootballBERT Configuration"""
+
+from transformers import PretrainedConfig
+
+class FootballBERTConfig(PretrainedConfig):
+    """
+    Configuration class for FootballBERT models.
+    
+    FootballBERT treats players as tokens and matches as sentences, learning
+    contextual embeddings for football players through masked player prediction.
+    
+    Args:
+        vocab_size (`int`, *optional*, defaults to 99944):
+            Vocabulary size of the FootballBERT model (number of unique players).
+        hidden_size (`int`, *optional*, defaults to 256):
+            Dimensionality of the encoder layers and the pooler layer.
+        num_hidden_layers (`int`, *optional*, defaults to 6):
+            Number of hidden layers in the Transformer encoder.
+        num_attention_heads (`int`, *optional*, defaults to 1):
+            Number of attention heads for each attention layer in the Transformer encoder.
+        intermediate_size (`int`, *optional*, defaults to 1024):
+            Dimensionality of the "intermediate" (often named feed-forward) layer in the Transformer encoder.
+        hidden_dropout_prob (`float`, *optional*, defaults to 0.05):
+            The dropout probability for all fully connected layers in the embeddings, encoder, and pooler.
+        max_position_embeddings (`int`, *optional*, defaults to 32):
+            The maximum sequence length (2 teams * 16 players max).
+        num_positions (`int`, *optional*, defaults to 1419):
+            Number of unique position labels.
+        num_seasons (`int`, *optional*, defaults to 23):
+            Number of unique seasons in the dataset.
+        pad_token_id (`int`, *optional*, defaults to 99944):
+            The id of the padding token.
+        mask_token_id (`int`, *optional*, defaults to 99943):
+            The id of the mask token.
+        position_pad_token_id (`int`, *optional*, defaults to 1419):
+            The id of the padding token for positions.
+        season_pad_token_id (`int`, *optional*, defaults to 23):
+            The id of the padding token for seasons.
+    """
+    
+    model_type = "footballbert"
+    
+    def __init__(
+        self,
+        vocab_size: int = 99944,
+        hidden_size: int = 256,
+        num_hidden_layers: int = 6,
+        num_attention_heads: int = 1,
+        intermediate_size: int = 1024,
+        hidden_dropout_prob: float = 0.05,
+        max_position_embeddings: int = 32,
+        num_positions: int = 1419,
+        num_seasons: int = 23,
+        pad_token_id: int = 99944,
+        mask_token_id: int = 99943,
+        position_pad_token_id: int = 1419,
+        season_pad_token_id: int = 23,
+        **kwargs
+    ):
+        super().__init__(pad_token_id=pad_token_id, **kwargs)
+        
+        self.vocab_size = vocab_size
+        self.hidden_size = hidden_size
+        self.num_hidden_layers = num_hidden_layers
+        self.num_attention_heads = num_attention_heads
+        self.intermediate_size = intermediate_size
+        self.hidden_dropout_prob = hidden_dropout_prob
+        self.max_position_embeddings = max_position_embeddings
+        self.num_positions = num_positions
+        self.num_seasons = num_seasons
+        self.mask_token_id = mask_token_id
+        self.position_pad_token_id = position_pad_token_id
+        self.season_pad_token_id = season_pad_token_id
+
+        # auto mapping for model
+        self.auto_map = {
+            "AutoConfig": "configuration_footballbert.FootballBERTConfig",
+            "AutoModel": "modeling_footballbert.FootballBERTModel",
+            "AutoModelForMaskedLM": "modeling_footballbert.FootballBERTForMaskedPlayerPrediction",
         }