Update

laithsakka · laithsakka · commit 465c89a7dca4 · 2025-09-03T15:30:39.000-07:00
[ghstack-poisoned]
diff --git a/torch/_meta_registrations.py b/torch/_meta_registrations.py
@@ -2376,7 +2376,7 @@ def _formula_transposed(ln: int, p: int, d: int, k: int, s: int, op: int) -> int
     from torch.fx.experimental.symbolic_shapes import sym_or
 
     torch._check(
-        sym_or(*(x > 0 for x in ret_shape[2:])),
+        sym_or(*[x > 0 for x in ret_shape[2:]]),
         lambda: f"Given input size per channel: {list(dims)}. "
         f"Calculated output size per channel: {ret_shape[2:]}. "
         f"Output size is too small",
