Merge pull request #1 from ankandrew/master

Minor Frozen GPTConfig

model.py

@@ -90,7 +90,7 @@ class Block(nn.Module):
         x = x + self.mlp(self.ln_2(x))
         return x
 
-@dataclass
+@dataclass(frozen=True)
 class GPTConfig:
     block_size: int = 1024
     vocab_size: int = 50257