Update modeling_plamo.py (#7)

- Update modeling_plamo.py (a071714cb4394ca127fb0282cb18d026b36050f5)


Co-authored-by: shintarou okada <[email protected]>

modeling_plamo.py

@@ -1434,7 +1434,7 @@ class Plamo2Model(Plamo2PreTrainedModel):
         require_attn_mask = False
         if not self.training or past_key_values is not None:
             require_attn_mask = True
-        if seq_length_with_past >= self.config.attention_window_size:
+        if seq_length_with_past > self.config.attention_window_size + 1:
             require_attn_mask = True
         if require_attn_mask and attention_mask is None:
             attention_mask = torch.ones(
@@ -1704,4 +1704,4 @@ class Bias(nn.Module):
         self,
         x: torch.Tensor,
     ) -> torch.Tensor:
-        return x + self._bias
+        return x + self._bias