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:
@@ -195,8 +195,6 @@ class Glm4MoeMTP(nn.Module, Glm4MixtureOfExperts):
|
||||
vllm_config=vllm_config, prefix=maybe_prefix(prefix, "model")
|
||||
)
|
||||
|
||||
self.expert_weights = []
|
||||
|
||||
# Set MoE hyperparameters
|
||||
self.num_moe_layers = self.config.num_nextn_predict_layers
|
||||
self.num_expert_groups = self.config.n_group
|
||||
|
||||
Reference in New Issue
Block a user