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 11, 17:30 UTC
[BugFix] Fix `AttributeError: 'MergedColumnParallelLinear' object has no attribute 'weight_scale'` (#30399)
Signed-off-by: Lucas Wilkinson <lwilkins@redhat.com>
Lucas Wilkinson committed
Dec 10, 2025 at 10:59 UTC
aacf0abf8bc219211b888a82f11f028e67b59531
1 file changed
+1
-1
vllm/model_executor/warmup/deep_gemm_warmup.py
+1
-1
@@ -89,7 +89,7 @@ def _extract_data_from_linear_base_module(
89
assert m.quant_method.quant_config is not None
90
91
w = m.weight
92
- ws = m.weight_scale
92
+ ws = m.weight_scale_inv if hasattr(m, "weight_scale_inv") else m.weight_scale
93
quant_block_size = m.quant_method.quant_config.weight_block_size
94
95
assert isinstance(w, torch.Tensor)