A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 4, 16:15 UTC
[Bugfix][CI] Fix wrong residual shape in TestFusedAddRMSNorm.example_inputs that causes flaky test (#40629)
Signed-off-by: Zhang Jian <jianmusings@gmail.com>
Zhang Jian committed
Apr 25, 2026 at 04:40 UTC
88256082058fdbd41281c4f1f9a19663a4d7a668
1 file changed
+3
-3
tests/compile/passes/test_functionalization.py
+3
-3
@@ -117,9 +117,9 @@ class TestFusedAddRMSNorm(torch.nn.Module):
117
else:
118
return norm_output, residual_output
119
120
- def example_inputs(self, batch_size=8, hidden_size=16, seq_len=16):
121
- hidden_states = torch.randn((batch_size * seq_len, hidden_size))
122
- residual = torch.randn((batch_size * seq_len, hidden_size))
120
+ def example_inputs(self, batch_size=8, seq_len=16):
121
+ hidden_states = torch.randn((batch_size * seq_len, self.hidden_size))
122
+ residual = torch.randn((batch_size * seq_len, self.intermediate_size))
123
return (hidden_states, residual)
124
125
def ops_in_model(self, do_fusion):