Update modeling_auristream.py

modeling_auristream.py

@@ -425,8 +425,8 @@ class CausalSelfAttention(nn.Module):
         self.c_attn = nn.Linear(self.n_embd, 3 * self.n_embd, bias=False)
         # output projection
         self.c_proj = nn.Linear(self.n_embd, self.n_embd, bias=False)
-
-        rope_theta = 500000
+        
+        rope_theta = 10000
         if hasattr(config, 'rope_theta') and config.rope_theta is not None:
             rope_theta = config.rope_theta