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:
@@ -217,7 +217,6 @@ class DeepSeekMTP(nn.Module, DeepseekV2MixtureOfExperts):
|
||||
self.set_moe_parameters()
|
||||
|
||||
def set_moe_parameters(self):
|
||||
self.expert_weights = []
|
||||
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