vllm kvprune wo:v1.1.0

d29c39ca · chenzk · f81ce56b · d29c39ca · d29c39ca · d29c39ca
Commit d29c39ca authored Apr 30, 2026 by chenzk
20 changed files
--- a/vllm/kvprune_legacy_save/triton_kernels/tensor.py
+++ b/vllm/kvprune_legacy_save/triton_kernels/tensor.py
--- a/vllm/kvprune_legacy_save/triton_kernels/tensor_details/__init__.py
+++ b/vllm/kvprune_legacy_save/triton_kernels/tensor_details/__init__.py
--- a/vllm/kvprune_legacy_save/triton_kernels/tensor_details/layout.py
+++ b/vllm/kvprune_legacy_save/triton_kernels/tensor_details/layout.py
--- a/vllm/kvprune_legacy_save/triton_kernels/tensor_details/layout_details/__init__.py
+++ b/vllm/kvprune_legacy_save/triton_kernels/tensor_details/layout_details/__init__.py
--- a/vllm/kvprune_legacy_save/triton_kernels/tensor_details/layout_details/base.py
+++ b/vllm/kvprune_legacy_save/triton_kernels/tensor_details/layout_details/base.py
--- a/vllm/kvprune_legacy_save/triton_kernels/tensor_details/layout_details/blackwell_scale.py
+++ b/vllm/kvprune_legacy_save/triton_kernels/tensor_details/layout_details/blackwell_scale.py
--- a/vllm/kvprune_legacy_save/triton_kernels/tensor_details/layout_details/blackwell_value.py
+++ b/vllm/kvprune_legacy_save/triton_kernels/tensor_details/layout_details/blackwell_value.py
--- a/vllm/kvprune_legacy_save/triton_kernels/tensor_details/layout_details/cdna4_scale.py
+++ b/vllm/kvprune_legacy_save/triton_kernels/tensor_details/layout_details/cdna4_scale.py
--- a/vllm/kvprune_legacy_save/triton_kernels/tensor_details/layout_details/hopper_scale.py
+++ b/vllm/kvprune_legacy_save/triton_kernels/tensor_details/layout_details/hopper_scale.py
--- a/vllm/kvprune_legacy_save/triton_kernels/tensor_details/layout_details/hopper_value.py
+++ b/vllm/kvprune_legacy_save/triton_kernels/tensor_details/layout_details/hopper_value.py
--- a/vllm/kvprune_legacy_save/triton_kernels/tensor_details/layout_details/strided.py
+++ b/vllm/kvprune_legacy_save/triton_kernels/tensor_details/layout_details/strided.py
--- a/vllm/kvprune_legacy_save/triton_kernels/testing.py
+++ b/vllm/kvprune_legacy_save/triton_kernels/testing.py
--- a/vllm/kvprune_legacy_save/triton_kernels/topk.py
+++ b/vllm/kvprune_legacy_save/triton_kernels/topk.py
--- a/vllm/kvprune_legacy_save/triton_kernels/topk_details/__init__.py
+++ b/vllm/kvprune_legacy_save/triton_kernels/topk_details/__init__.py
--- a/vllm/kvprune_legacy_save/triton_kernels/topk_details/_topk_backward.py
+++ b/vllm/kvprune_legacy_save/triton_kernels/topk_details/_topk_backward.py
--- a/vllm/kvprune_legacy_save/triton_kernels/topk_details/_topk_forward.py
+++ b/vllm/kvprune_legacy_save/triton_kernels/topk_details/_topk_forward.py
--- a/vllm/kvprune_legacy_save/utils/__init__.py
+++ b/vllm/kvprune_legacy_save/utils/__init__.py
--- a/vllm/kvprune_legacy_save/utils/arguments.py
+++ b/vllm/kvprune_legacy_save/utils/arguments.py
--- a/vllm/kvprune_legacy_save/utils/context.py
+++ b/vllm/kvprune_legacy_save/utils/context.py
--- a/vllm/kvprune_legacy_save/utils/helpers.py
+++ b/vllm/kvprune_legacy_save/utils/helpers.py