Update configuration_helpingai.py

configuration_helpingai.py

@@ -1,61 +1,3 @@
-from transformers.configuration_utils import PretrainedConfig
-
-
-class HelpingAIConfig(PretrainedConfig):
-    model_type = "helpingai"
-
-    def __init__(
-        self,
-        vocab_size=50257,
-        hidden_size=768,
-        num_hidden_layers=12,
-        num_attention_heads=12,
-        intermediate_size=3072,
-        max_position_embeddings=2048,
-        layer_norm_epsilon=1e-5,
-        hidden_act="gelu",
-        dropout=0.0,
-        attention_dropout=0.0,
-        tie_word_embeddings=True,
-        # Structured output head
-        use_structured_output=True,
-        structured_output_vocab_size=2,
-        # Speech head
-        use_speech_output=False,
-        speech_num_mels=80,
-        speech_head_hidden_dim=1024,
-        speech_upsample_factor=1,
-        speech_loss_type="l1",
-        # Misc
-        initializer_range=0.02,
-        **kwargs,
-    ):
-        super().__init__(tie_word_embeddings=tie_word_embeddings, **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.max_position_embeddings = max_position_embeddings
-        self.layer_norm_epsilon = layer_norm_epsilon
-        self.hidden_act = hidden_act
-        self.dropout = dropout
-        self.attention_dropout = attention_dropout
-        self.initializer_range = initializer_range
-
-        # Structured
-        self.use_structured_output = use_structured_output
-        self.structured_output_vocab_size = structured_output_vocab_size
-
-        # Speech
-        self.use_speech_output = use_speech_output
-        self.speech_num_mels = speech_num_mels
-        self.speech_head_hidden_dim = speech_head_hidden_dim
-        self.speech_upsample_factor = speech_upsample_factor
-        self.speech_loss_type = speech_loss_type
-
-"""HelpingAI model configuration"""
-
 from transformers.configuration_utils import PretrainedConfig, layer_type_validation
 from transformers.modeling_rope_utils import rope_config_validation
 from transformers.utils import logging
@@ -264,10 +206,10 @@ class HelpingAIConfig(PretrainedConfig):
         structured_output_vocab_size=100,
         empathy_scaling_factor=1.2,
         reasoning_temperature=0.8,
-    # Structured head architecture (new)
-    structured_head_type: str = "linear",  # one of: linear, mlp_v1
-    structured_head_hidden_dim: int | None = None,
-    structured_head_activation: str = "gelu",  # gelu or relu
+        # Structured head architecture (new)
+        structured_head_type: str = "linear",  # one of: linear, mlp_v1
+        structured_head_hidden_dim: int | None = None,
+        structured_head_activation: str = "gelu",  # gelu or relu
         # Speech output head options
         use_speech_output=False,
         speech_num_mels=80,