[Feat] Support PP with PCP in GPU Model Runner V2 (#59139)

Signed-off-by: QiuChunshuo <qiuchunshuo@huawei.com>
Co-authored-by: Lucas Wilkinson <LucasWilkinson@users.noreply.github.com>
This commit is contained in:
Qiu Chunshuo
2026-09-30 10:21:56 -04:00
committed by GitHub
co-authored by Lucas Wilkinson
parent 2e52a558f0
commit 98b45c1a94
4 changed files with 46 additions and 3 deletions
+2
View File
@@ -2064,6 +2064,8 @@ class GPUModelRunner(LoRAModelRunnerMixin):
# Non-last PP rank: hidden_states is None because this rank produced
# IntermediateTensors instead of final hidden states. Receive the
# sampled tokens broadcast from the last rank and update local state.
if self.pcp_manager is not None:
input_batch = self.pcp_manager.global_batch
assert self.pp_handler is not None
all_decode_next = self.pp_handler.receive(input_batch)
# Optimistically update num_computed_tokens for entire batch here.