A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 7, 21:30 UTC
[Mamba1] - Change supports_update_block_table to True (#35054)
Signed-off-by: Josephasafg <ajgard7@gmail.com>
Asaf Gardin committed
Feb 24, 2026 at 05:05 UTC
95642441d0e72f7633f18da053c9e7a9862a22c2
1 file changed
-1
vllm/v1/attention/backends/mamba1_attn.py
-1
@@ -29,4 +29,3 @@ class Mamba1AttentionMetadataBuilder(
29
BaseMambaAttentionMetadataBuilder[Mamba1AttentionMetadata]
30
):
31
metadata_cls = Mamba1AttentionMetadata
32
- supports_update_block_table: bool = False