[NPU] fix pp_size>1 (#12195)
This commit is contained in:
@@ -1659,9 +1659,11 @@ class ModelRunner:
|
||||
get_attention_tp_size()
|
||||
),
|
||||
head_dim=self.model_config.head_dim,
|
||||
layer_num=self.model_config.num_hidden_layers,
|
||||
layer_num=self.num_effective_layers,
|
||||
device=self.device,
|
||||
enable_memory_saver=self.server_args.enable_memory_saver,
|
||||
start_layer=self.start_layer,
|
||||
end_layer=self.end_layer,
|
||||
)
|
||||
elif self.use_mla_backend and is_nsa_model:
|
||||
self.token_to_kv_pool = NSATokenToKVPool(
|
||||
|
||||
Reference in New Issue
Block a user