Fix Transformers backend FP8 MoE and remove some boilerplate (#46820)

Signed-off-by: Harry Mellor <19981378+hmellor@users.noreply.github.com>
This commit is contained in:
Harry Mellor
2026-06-27 00:16:05 +01:00
committed by GitHub
parent 2ff76a5e85
commit d8eb734d94
29 changed files with 80 additions and 156 deletions
-1
View File
@@ -1562,7 +1562,6 @@ class Gemma4ForCausalLM(
)
# --- MixtureOfExperts protocol ---
self.expert_weights: list[list[torch.Tensor]] = []
self.moe_layers: list[nn.Module] = []
example_moe: Gemma4MoE | None = None