Add online current-tail LoRA benchmark

omni_cached_av.py

@@ -192,6 +192,7 @@ def reconstruct_open_stream_sequence(
     store: CachedBlockStore,
     static: dict[str, Any],
     anchor: dict[str, Any],
+    current_piece: dict[str, torch.Tensor] | None = None,
 ) -> tuple[torch.Tensor, torch.LongTensor]:
     """Materialize one differentiable Thinker input from frozen AV embeddings.
 
@@ -219,7 +220,13 @@ def reconstruct_open_stream_sequence(
     last_shift = 0
 
     for sequence_index, (block_index, tail_steps) in enumerate(selected):
-        piece = store.variant(block_index, tail_steps)
+        is_current = sequence_index == len(selected) - 1
+        piece = (
+            current_piece
+            if is_current and current_piece is not None
+            else store.variant(block_index, tail_steps)
+        )
+        validate_piece(piece)
         local_payload_positions = _strip_position_batch(piece["payload_position_ids"])
         local_payload_base = int(local_payload_positions.min().item())
         if first_payload_base is None: