configuration_easyformer.py

# configuration_easyformer.py
from transformers import PretrainedConfig

class EasyFormerConfig(PretrainedConfig):
    model_type = "transformer"

    def __init__(
        self,
        vocab_size=8000,
        d_model=192,
        n_layer=5,
        n_head=4,
        ctx=128,
        dropout=0.1,
        **kwargs,
    ):
        super().__init__(**kwargs)
        self.vocab_size = vocab_size
        self.d_model = d_model
        self.n_layer = n_layer
        self.n_head = n_head
        self.ctx = ctx
        self.dropout = dropout

    # --- HF-совместимые property ---
    @property
    def num_hidden_layers(self):
        return self.n_layer

    @property
    def hidden_size(self):
        return self.d_model

    @property
    def num_attention_heads(self):
        return self.n_head

    @property
    def intermediate_size(self):
        return 2 * self.d_model

    @property
    def max_position_embeddings(self):
        return self.ctx

    @property
    def num_key_value_heads(self):
        return self.n_head

    @property
    def head_dim(self):
        return self.d_model // max(1, self.n_head)