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
@@ -134,7 +134,6 @@ class GlmOcrMTP(nn.Module, SupportsPP):
vllm_config=vllm_config, prefix=maybe_prefix(prefix, "model")
)
self.expert_weights = []
self.num_layers = self.config.num_nextn_predict_layers
for layer in self.model.layers.values():
assert isinstance(layer, GlmOcrMultiTokenPredictorLayer)