diff --git a/vall_e/vall_e/base.py b/vall_e/vall_e/base.py
index da706b2..bd9fe36 100644
--- a/vall_e/vall_e/base.py
+++ b/vall_e/vall_e/base.py
@@ -231,6 +231,10 @@ class Base(nn.Module):
     def use_stop_token(self) -> bool:
         raise NotImplementedError
 
+    @property
+    def n_prom_levels(self) -> int:
+        return 8
+
     def __init__(
         self,
         n_tokens: int,
@@ -238,7 +242,6 @@ class Base(nn.Module):
         n_heads: int = 8,
         n_layers: int = 12,
         p_dropout: float = 0.1,
-        n_prom_levels: int = 8,
         resp_loss_only: bool = False,
     ):
         super().__init__()
@@ -254,7 +257,7 @@ class Base(nn.Module):
 
         # It's not clear whether the whole prom are used or only the first level quantization
         # Just use all of them as it is more sufficient and we don't need to sample it, or do we?
-        self.prom_emb = MultiEmbedding(n_tokens, d_model, n_levels=n_prom_levels)
+        self.prom_emb = MultiEmbedding(n_tokens, d_model, n_levels=self.n_prom_levels)
 
         # +1 to include the stop token
         # Note that, for different levels, I don't use AdaLN for simplicity