Skip to content

[ET-VK][Ops] dequantize_per_tensor.tensor variant #12209

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 10 commits into from
Jul 14, 2025
Merged

[ET-VK][Ops] dequantize_per_tensor.tensor variant #12209

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
10 commits
Select commit Hold shift + click to select a range
10512c5
[ET-VK][Ops] dequantize_per_tensor.tensor variant
Jul 3, 2025
83b10f7
Update on "[ET-VK][Ops] dequantize_per_tensor.tensor variant"
Jul 3, 2025
32833a6
Update on "[ET-VK][Ops] dequantize_per_tensor.tensor variant"
Jul 7, 2025
bcd6d55
Update on "[ET-VK][Ops] dequantize_per_tensor.tensor variant"
Jul 10, 2025
d9c3310
Update on "[ET-VK][Ops] dequantize_per_tensor.tensor variant"
Jul 11, 2025
260376f
Update on "[ET-VK][Ops] dequantize_per_tensor.tensor variant"
Jul 11, 2025
87bf615
Update on "[ET-VK][Ops] dequantize_per_tensor.tensor variant"
Jul 11, 2025
a322267
Update on "[ET-VK][Ops] dequantize_per_tensor.tensor variant"
Jul 11, 2025
0990ef8
Update on "[ET-VK][Ops] dequantize_per_tensor.tensor variant"
Jul 11, 2025
96c1a96
Update on "[ET-VK][Ops] dequantize_per_tensor.tensor variant"
Jul 14, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
7 changes: 4 additions & 3 deletions backends/vulkan/runtime/graph/ops/glsl/dequantize_buffer.glsl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -27,9 +27,10 @@ ${layout_declare_tensor(B, "w", "t_out", OUT_DTYPE, "buffer")}
${layout_declare_tensor(B, "r", "t_in", IN_DTYPE, "buffer")}

$if MODE == "per_tensor":
${layout_declare_tensor(B, "r", "t_scale", "float", "buffer")}
${layout_declare_tensor(B, "r", "t_zero_point", "int", "buffer")}

layout(push_constant) uniform restrict Block {
float scale;
int zero_point;
int quant_min;
int quant_max;
};
Expand Down Expand Up @@ -146,7 +147,7 @@ void dequantize_per_tensor() {
const int in_bufi = tidx_to_bufi(out_tidx, t_in_strides);

IN_T qvalue = t_in[in_bufi];
OUT_T value = dequantize_val(qvalue, scale, zero_point);
OUT_T value = dequantize_val(qvalue, t_scale[0], t_zero_point[0]);

t_out[out_bufi] = value;
}
Expand Down
8 changes: 5 additions & 3 deletions backends/vulkan/runtime/graph/ops/glsl/dequantize_texture.glsl
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -30,9 +30,10 @@ ${layout_declare_tensor(B, "w", "t_out", OUT_DTYPE, "texture3d")}
${layout_declare_tensor(B, "r", "t_in", IN_DTYPE, "texture3d")}

$if MODE == "per_tensor":
${layout_declare_tensor(B, "r", "t_scale", "float", "buffer")}
${layout_declare_tensor(B, "r", "t_zero_point", "int", "buffer")}

layout(push_constant) uniform restrict Block {
float scale;
int zero_point;
int quant_min;
int quant_max;
};
Expand Down Expand Up @@ -148,7 +149,8 @@ void dequantize_per_tensor() {

[[unroll]] for (int i = 0; i < 4; ++i) {
IN_T qvalue = IN_T(intex[i]);
OUT_T value = dequantize_val(qvalue, scale, zero_point);
OUT_T value = dequantize_val(qvalue, t_scale[0], t_zero_point[0]);

$if OUT_DTYPE == "double":
outtex[i] = float(value);
$else:
Expand Down
25 changes: 9 additions & 16 deletions backends/vulkan/runtime/graph/ops/impl/Dequantize.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -85,8 +85,6 @@ void add_dequantize_per_tensor_node(
add_dtype_suffix(kernel_name, graph.dtype_of(input));
add_dtype_suffix(kernel_name, graph.dtype_of(output));

float scale_val = static_cast<float>(graph.get_double(scale));
int zero_point_val = static_cast<int>(graph.get_int(zero_point));
int quant_min_val = static_cast<int>(graph.get_int(quant_min));
int quant_max_val = static_cast<int>(graph.get_int(quant_max));

Expand All @@ -100,23 +98,16 @@ void add_dequantize_per_tensor_node(
graph.strides_ubo(input),
graph.sizes_ubo(output),
graph.strides_ubo(output)};
push_constants = {
PushConstantDataInfo(&scale_val, sizeof(float)),
PushConstantDataInfo(&zero_point_val, sizeof(int)),
PushConstantDataInfo(&quant_min_val, sizeof(int)),
PushConstantDataInfo(&quant_max_val, sizeof(int)),
};
} else {
param_ubos = {
graph.logical_limits_ubo(input), graph.logical_limits_ubo(output)};
push_constants = {
PushConstantDataInfo(&scale_val, sizeof(float)),
PushConstantDataInfo(&zero_point_val, sizeof(int)),
PushConstantDataInfo(&quant_min_val, sizeof(int)),
PushConstantDataInfo(&quant_max_val, sizeof(int)),
};
}

push_constants = {
PushConstantDataInfo(&quant_min_val, sizeof(int)),
PushConstantDataInfo(&quant_max_val, sizeof(int)),
};

vkapi::SpecVarList spec_vars = {
graph.hashed_layout_of(output),
graph.hashed_layout_of(input),
Expand All @@ -128,7 +119,9 @@ void add_dequantize_per_tensor_node(
default_pick_global_wg_size,
default_pick_local_wg_size,
// Inputs and Outputs
{{output, vkapi::kWrite}, {input, vkapi::kRead}},
{{output, vkapi::kWrite},
{input, vkapi::kRead},
{{scale, zero_point}, vkapi::kRead}},
// Shader param buffers
param_ubos,
// Push Constants
Expand Down Expand Up @@ -517,7 +510,7 @@ void dequantize_per_channel_impl(

REGISTER_OPERATORS {
VK_REGISTER_OP(
quantized_decomposed.dequantize_per_tensor.default,
quantized_decomposed.dequantize_per_tensor.tensor,
dequantize_per_tensor_impl);
VK_REGISTER_OP(
quantized_decomposed.dequantize_per_token.default,
Expand Down
Loading
Loading