Merge tag 'v0.14.0' into v0.14.0-dev

7e63ef82 · zhuwenwen · 8cbcac5d · b17039bc · 7e63ef82 · 7e63ef82
Commit 7e63ef82 authored Jan 21, 2026 by zhuwenwen
20 changed files
--- a/vllm/model_executor/layers/activation.py
+++ b/vllm/model_executor/layers/activation.py
--- a/vllm/model_executor/layers/attention/__init__.py
+++ b/vllm/model_executor/layers/attention/__init__.py
--- a/vllm/attention/layers/chunked_local_attention.py
+++ b/vllm/attention/layers/chunked_local_attention.py
--- a/vllm/attention/layers/cross_attention.py
+++ b/vllm/attention/layers/cross_attention.py
--- a/vllm/attention/layers/encoder_only_attention.py
+++ b/vllm/attention/layers/encoder_only_attention.py
--- a/vllm/attention/layers/mm_encoder_attention.py
+++ b/vllm/attention/layers/mm_encoder_attention.py
--- a/vllm/model_executor/layers/attention/static_sink_attention.py
+++ b/vllm/model_executor/layers/attention/static_sink_attention.py
--- a/vllm/model_executor/layers/attention_layer_base.py
+++ b/vllm/model_executor/layers/attention_layer_base.py
--- a/vllm/model_executor/layers/batch_invariant.py
+++ b/vllm/model_executor/layers/batch_invariant.py
--- a/vllm/model_executor/layers/conv.py
+++ b/vllm/model_executor/layers/conv.py
--- a/vllm/model_executor/layers/fla/ops/fused_recurrent.py
+++ b/vllm/model_executor/layers/fla/ops/fused_recurrent.py
--- a/vllm/model_executor/layers/fla/ops/utils.py
+++ b/vllm/model_executor/layers/fla/ops/utils.py
--- a/vllm/model_executor/layers/fused_moe/__init__.py
+++ b/vllm/model_executor/layers/fused_moe/__init__.py
--- a/vllm/model_executor/layers/fused_moe/all2all_utils.py
+++ b/vllm/model_executor/layers/fused_moe/all2all_utils.py
--- a/vllm/model_executor/layers/fused_moe/batched_deep_gemm_moe.py
+++ b/vllm/model_executor/layers/fused_moe/batched_deep_gemm_moe.py
--- a/vllm/model_executor/layers/fused_moe/config.py
+++ b/vllm/model_executor/layers/fused_moe/config.py
--- a/vllm/model_executor/layers/fused_moe/configs/E=128,N=192,device_name=gfx938_64cu.json
+++ b/vllm/model_executor/layers/fused_moe/configs/E=128,N=192,device_name=gfx938_64cu.json
--- a/vllm/model_executor/layers/fused_moe/configs/E=128,N=192,device_name=gfx938_64cu_nn.json
+++ b/vllm/model_executor/layers/fused_moe/configs/E=128,N=192,device_name=gfx938_64cu_nn.json
--- a/vllm/model_executor/layers/fused_moe/configs/E=128,N=512,device_name=NVIDIA_B200,dtype=fp8_w8a8,block_shape=[128,128].json
+++ b/vllm/model_executor/layers/fused_moe/configs/E=128,N=512,device_name=NVIDIA_B200,dtype=fp8_w8a8,block_shape=[128,128].json
--- a/vllm/model_executor/layers/fused_moe/configs/E=128,N=704,device_name=NVIDIA_RTX_PRO_6000_Blackwell_Workstation_Edition,dtype=fp8_w8a8.json
+++ b/vllm/model_executor/layers/fused_moe/configs/E=128,N=704,device_name=NVIDIA_RTX_PRO_6000_Blackwell_Workstation_Edition,dtype=fp8_w8a8.json