[Bugfix] Fix RMS norm + quant fusion on DeepGEMM UE8M0 path for B200 (#40552)

Signed-off-by: Lucas Kabela <lucaskabela@meta.com>
This commit is contained in:
Lucas Kabela
2026-04-22 22:04:42 +00:00
committed by GitHub
parent 9c271f9403
commit b8401a9bf4
2 changed files with 22 additions and 0 deletions
+1
View File
@@ -1826,6 +1826,7 @@ class TestFP8Layer(torch.nn.Module):
self.weight = torch.rand(weight_shape).to(dtype=FP8_DTYPE)
self.input_scale = None
self.weight_scale = None
self.weight_block_size = [block_size, block_size]
if transpose_weights:
self.weight = self.weight.t()
else: