zhujiangang commited on
Commit
c9c10d7
·
verified ·
1 Parent(s): 6a4f845

Update configuration_bailing_moe.py

Browse files
Files changed (1) hide show
  1. configuration_bailing_moe.py +2 -2
configuration_bailing_moe.py CHANGED
@@ -63,7 +63,7 @@ class BailingMoeConfig(PretrainedConfig):
63
  self.use_sliding_window = use_sliding_window
64
  self.sliding_window = sliding_window
65
  self.max_window_layers = max_window_layers
66
- self.head_dim = head_dim
67
  self.rope_scaling = rope_scaling
68
 
69
  # MoE configs
@@ -75,4 +75,4 @@ class BailingMoeConfig(PretrainedConfig):
75
  self.first_k_dense_replace = first_k_dense_replace
76
  self.output_router_logits = output_router_logits
77
 
78
- super().__init__(pad_token_id=pad_token_id, tie_word_embeddings=tie_word_embeddings, **kwargs)
 
63
  self.use_sliding_window = use_sliding_window
64
  self.sliding_window = sliding_window
65
  self.max_window_layers = max_window_layers
66
+ self.head_dim = head_dim or self.hidden_size // self.num_attention_heads
67
  self.rope_scaling = rope_scaling
68
 
69
  # MoE configs
 
75
  self.first_k_dense_replace = first_k_dense_replace
76
  self.output_router_logits = output_router_logits
77
 
78
+ super().__init__(pad_token_id=pad_token_id, tie_word_embeddings=tie_word_embeddings, **kwargs)