[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:
@@ -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:
|
||||
|
||||
Reference in New Issue
Block a user