Skip to content

Commit d15bdb5

Browse files
author
Nathanael See
committed
Update base for Update on "[ET-VK][int4] Wrap int4 linear calls with view_copy nodes to squeeze/unsqueeze inputs"
This is done automatically for full-precision linear/mm nodes in the graph at torch.export graph tracing time, but is not done for the int4 op. The new pass adds view_copy nodes, as there are subsequent passes which can fuse view_copy nodes if redundant, and convert view_copy nodes to squeeze/unsqueeze nodes. Differential Revision: [D69065866](https://our.internmc.facebook.com/intern/diff/D69065866/) [ghstack-poisoned]
1 parent 9873869 commit d15bdb5

File tree

1 file changed

+2
-2
lines changed

1 file changed

+2
-2
lines changed

backends/vulkan/runtime/graph/ops/impl/QuantizedLinear.cpp

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -352,8 +352,8 @@ void add_q_4w_linear_node(
352352
local_wg_size,
353353
// Inputs and Outputs
354354
{{out_W_packed, vkapi::MemoryAccessType::WRITE},
355-
{{mat1_W_packed, mat2, scales_and_zeros},
356-
vkapi::MemoryAccessType::READ}},
355+
{{mat1_W_packed, mat2, scales_and_zeros},
356+
vkapi::MemoryAccessType::READ}},
357357
// Shader params buffers
358358
ubos,
359359
// Specialization Constants

0 commit comments

Comments
 (0)