File size: 1,458 Bytes
af0d032 dd334bd af0d032 dd334bd af0d032 dd334bd af0d032 dd334bd af0d032 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 | from transformers import PretrainedConfig
QWEN_TOKEN_ID = 151643
AUTO_MAP = {
"AutoConfig": "configuration_modern_llm.ModernLLMConfig",
"AutoModelForCausalLM": "modeling_modern_llm.ModernLLMForCausalLM",
}
class ModernLLMConfig(PretrainedConfig):
model_type = "modern_llm"
def __init__(
self,
vocab_size: int = 151936,
hidden_size: int = 768,
intermediate_size: int = 2048,
num_hidden_layers: int = 12,
num_attention_heads: int = 12,
num_key_value_heads: int = 4,
max_position_embeddings: int = 2048,
rms_norm_eps: float = 1e-6,
rope_theta: float = 1000000.0,
pad_token_id: int = QWEN_TOKEN_ID,
bos_token_id: int = QWEN_TOKEN_ID,
eos_token_id: int = QWEN_TOKEN_ID,
**kwargs,
):
self.vocab_size = vocab_size
self.hidden_size = hidden_size
self.intermediate_size = intermediate_size
self.num_hidden_layers = num_hidden_layers
self.num_attention_heads = num_attention_heads
self.num_key_value_heads = num_key_value_heads
self.max_position_embeddings = max_position_embeddings
self.rms_norm_eps = rms_norm_eps
self.rope_theta = rope_theta
self.auto_map = dict(AUTO_MAP)
super().__init__(
pad_token_id=pad_token_id,
bos_token_id=bos_token_id,
eos_token_id=eos_token_id,
**kwargs,
)
|