Update configuration_telechat.py

configuration_telechat.py

@@ -119,6 +119,7 @@ class TELECHATConfig(PretrainedConfig):
             n_embd=768,
             n_layer=12,
             n_head=12,
+            num_key_value_heads=64,
             n_inner=None,
             activation_function="gelu_new",
             resid_pdrop=0.1,
@@ -159,6 +160,7 @@ class TELECHATConfig(PretrainedConfig):
         self.n_embd = n_embd
         self.n_layer = n_layer
         self.n_head = n_head
+        self.num_key_value_heads = num_key_value_heads
         self.n_inner = n_inner
         self.activation_function = activation_function
         self.resid_pdrop = resid_pdrop