| """ HelpingAI model configuration""" |
|
|
| from transformers import PretrainedConfig |
| from transformers.utils import logging |
|
|
|
|
| logger = logging.get_logger(__name__) |
|
|
| class HelpingAIConfig(PretrainedConfig): |
| keys_to_ignore_at_inference = ["past_key_values"] |
| model_type = "HelpingAI" |
| def __init__( |
| self, |
| vocab_size=50304, |
| hidden_size=2560, |
| intermediate_size=6912, |
| num_hidden_layers=32, |
| num_attention_heads=32, |
| num_key_value_heads=32, |
| head_dim=256, |
| hidden_act="silu", |
| max_position_embeddings=4096, |
| initializer_range=0.02, |
| rms_norm_eps=1e-6, |
| use_cache=True, |
| hidden_activation=None, |
| rope_theta=10000, |
| rope_pct=0.25, |
| attention_bias=False, |
| attention_dropout=0.0, |
| num_experts_per_tok=2, |
| num_local_experts=8, |
| router_aux_loss_coef=0.02, |
| output_router_logits=False, |
| norm_eps=1.0e-5, |
| **kwargs, |
| ): |
| self.vocab_size = vocab_size |
| self.max_position_embeddings = max_position_embeddings |
| 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.head_dim = head_dim |
| self.hidden_act = hidden_act |
| self.hidden_activation = hidden_activation |
| self.num_key_value_heads = num_key_value_heads |
| self.initializer_range = initializer_range |
| self.rms_norm_eps = rms_norm_eps |
| self.use_cache = use_cache |
| self.rope_theta = rope_theta |
| self.attention_bias = attention_bias |
| self.attention_dropout = attention_dropout |
| self.num_experts_per_tok = num_experts_per_tok |
| self.num_local_experts = num_local_experts |
| self.router_aux_loss_coef = router_aux_loss_coef |
| self.output_router_logits = output_router_logits |
| self.rope_pct = rope_pct |
| self.norm_eps = norm_eps |
| super().__init__(**kwargs) |
|
|
