Deepseek-v3 Batch Invariant on 8xH100 (#26609)

Signed-off-by: Bram Wasti <bwasti@meta.com>
Co-authored-by: Wentao Ye <44945378+yewentao256@users.noreply.github.com>
This commit is contained in:
Bram Wasti
2025-10-15 22:06:02 -07:00
committed by GitHub
co-authored by Wentao Ye
parent 785d8b6410
commit 7d8975de84
21 changed files with 1567 additions and 102 deletions
-3
View File
@@ -231,9 +231,6 @@ class GPUModelRunner(LoRAModelRunnerMixin, KVConnectorModelRunnerMixin):
from vllm.model_executor.models.utils import set_cpu_offload_max_bytes
set_cpu_offload_max_bytes(int(self.cache_config.cpu_offload_gb * 1024**3))
from vllm.model_executor.layers.batch_invariant import init_batch_invariance
init_batch_invariance()
model_config = self.model_config
cache_config = self.cache_config