From 7268457999d54f4014f2815911837c9277ca23f5 Mon Sep 17 00:00:00 2001 From: Varun Sundar Rabindranath Date: Wed, 3 Jun 2026 03:03:00 -0400 Subject: [PATCH] [KV Offloading] Enable HMA models for Tiering Offloading (#44287) Signed-off-by: varun sundar rabindranath Co-authored-by: varun sundar rabindranath --- vllm/v1/kv_offload/tiering/spec.py | 1 - 1 file changed, 1 deletion(-) diff --git a/vllm/v1/kv_offload/tiering/spec.py b/vllm/v1/kv_offload/tiering/spec.py index ced8a7fc65..8a2759440d 100644 --- a/vllm/v1/kv_offload/tiering/spec.py +++ b/vllm/v1/kv_offload/tiering/spec.py @@ -110,7 +110,6 @@ class TieringOffloadingSpec(CPUOffloadingSpec): self._scheduler_mmap = scheduler_mmap # Create primary tier (CPU-based) - assert len(self.gpu_block_size) == 1 primary_tier = CPUPrimaryTierOffloadingManager( num_blocks=self.num_blocks, cache_policy=self.eviction_policy, # type: ignore[arg-type]