Skip to content

[cherry-pick] fix conv+hard_swish in int8-int8 compute error #9001

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 1 commit into from
May 10, 2022
Merged

[cherry-pick] fix conv+hard_swish in int8-int8 compute error #9001

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
25 changes: 23 additions & 2 deletions lite/backends/arm/math/gemv_arm_int8.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1222,6 +1222,8 @@ inline void write_gemv_out(const int* in,
float32x4_t valpha = vdupq_n_f32(alpha);
float32x4_t voffset = vdupq_n_f32(offset);
float32x4_t vthreshold = vdupq_n_f32(threshold);
float32x4_t vmax = vdupq_n_f32(-127.f);

#ifdef __aarch64__
asm volatile(
"cmp %w[cnt], #1\n"
Expand Down Expand Up @@ -1252,6 +1254,11 @@ inline void write_gemv_out(const int* in,
"fmul v0.4s, v4.4s, v5.4s\n"
"fmin v6.4s, v6.4s, %[vthreshold].4s\n"
"fmul v3.4s, v6.4s, v7.4s\n"
// out >= -127
"fcmge v4.4s, v0.4s, %[vmax].4s\n"
"fcmge v5.4s, v3.4s, %[vmax].4s\n"
"bif v0.16b, %[vmax].16b, v4.16b\n"
"bif v3.16b, %[vmax].16b, v5.16b\n"
// fp32 - int32
"fcvtas v4.4s, v0.4s\n"
"fcvtas v5.4s, v3.4s\n"
Expand Down Expand Up @@ -1279,6 +1286,9 @@ inline void write_gemv_out(const int* in,
"fmax v4.4s, v4.4s, %[vzero].4s\n"
"fmin v4.4s, v4.4s, %[vthreshold].4s\n"
"fmul v0.4s, v4.4s, v5.4s\n"
// out >= -127
"fcmge v4.4s, v0.4s, %[vmax].4s\n"
"bif v0.16b, %[vmax].16b, v4.16b\n"
// fp32 - int32
"fcvtas v4.4s, v0.4s\n"
// int32 - int16
Expand All @@ -1298,7 +1308,8 @@ inline void write_gemv_out(const int* in,
[vzero] "w"(vzero),
[valpha] "w"(valpha),
[voffset] "w"(voffset),
[vthreshold] "w"(vthreshold)
[vthreshold] "w"(vthreshold),
[vmax] "w"(vmax)
: "cc",
"memory",
"v0",
Expand Down Expand Up @@ -1349,6 +1360,11 @@ inline void write_gemv_out(const int* in,
"vbif q13, %q[vfive], q10\n"
"vadd.f32 q5, q5, q12\n"
"vadd.f32 q8, q8, q13\n"
// data >= -127
"vcge.f32 q7, q5, %q[vmax]\n"
"vcge.f32 q9, q8, %q[vmax]\n"
"vbif q5, %q[vmax], q7\n"
"vbif q8, %q[vmax], q9\n"
// fp32 -> int32
"vcvt.s32.f32 q7, q5\n"
"vcvt.s32.f32 q9, q8\n"
Expand Down Expand Up @@ -1380,6 +1396,9 @@ inline void write_gemv_out(const int* in,
"vcge.f32 q7, q5, %q[vzero]\n"
"vbif q12, %q[vfive], q7\n"
"vadd.f32 q5, q5, q12\n"
// data >= -127
"vcge.f32 q7, q5, %q[vmax]\n"
"vbif q5, %q[vmax], q7\n"
// fp32 -> int32
"vcvt.s32.f32 q7, q5\n"
// int32 -> int16
Expand All @@ -1400,7 +1419,8 @@ inline void write_gemv_out(const int* in,
[valpha] "w"(valpha),
[voffset] "w"(voffset),
[vthreshold] "w"(vthreshold),
[vfive] "w"(vfive)
[vfive] "w"(vfive),
[vmax] "w"(vmax)
: "cc",
"memory",
"q4",
Expand Down Expand Up @@ -1624,6 +1644,7 @@ bool gemv_int8_trans_oth(const int8_t* A,
memset(zerobuf, 0, sizeof(float) * (M + 16));
const float* bias_ptr = is_bias ? bias : zerobuf;
float six = alpha;

#ifdef __aarch64__
int cnt = N >> 3;
int tail = N & 7;
Expand Down
1 change: 1 addition & 0 deletions lite/kernels/arm/conv_direct.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -154,6 +154,7 @@ inline bool direct_conv_trans_weights<PRECISION(kInt8), PRECISION(kInt8)>(
}
//! update hardswish parameter
if (act_param.active_type == lite_api::ActivationType::kHardSwish) {
act_param.hard_swish_scale = act_param.hard_swish_scale / out_scale;
act_param.hard_swish_offset = act_param.hard_swish_offset / out_scale;
act_param.hard_swish_threshold = act_param.hard_swish_threshold / out_scale;
}
Expand Down
2 changes: 2 additions & 0 deletions lite/kernels/arm/conv_gemmlike.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -86,6 +86,8 @@ void GemmLikeConv<PRECISION(kInt8), PRECISION(kInt8)>::PrepareForRun() {
//! update hardswish parameter
if (param.activation_param.active_type ==
lite_api::ActivationType::kHardSwish) {
param.activation_param.hard_swish_scale =
param.activation_param.hard_swish_scale / param.output_scale;
param.activation_param.hard_swish_offset =
param.activation_param.hard_swish_offset / param.output_scale;
param.activation_param.hard_swish_threshold =
Expand Down
2 changes: 2 additions & 0 deletions lite/kernels/arm/conv_transpose_compute.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -154,6 +154,8 @@ void Conv2DTransposeCompute<PRECISION(kInt8),
//! update hardswish parameter
if (param.activation_param.active_type ==
lite_api::ActivationType::kHardSwish) {
param.activation_param.hard_swish_scale =
param.activation_param.hard_swish_scale / param.output_scale;
param.activation_param.hard_swish_offset =
param.activation_param.hard_swish_offset / param.output_scale;
param.activation_param.hard_swish_threshold =
Expand Down
2 changes: 2 additions & 0 deletions lite/kernels/arm/conv_winograd.cc
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -196,6 +196,8 @@ void WinogradConv<PRECISION(kInt8), OutType>::ReInitWhenNeeded() {
//! update hardswish parameter
if (param.activation_param.active_type ==
lite_api::ActivationType::kHardSwish) {
param.activation_param.hard_swish_scale =
param.activation_param.hard_swish_scale / param.output_scale;
param.activation_param.hard_swish_offset =
param.activation_param.hard_swish_offset / output_scale;
param.activation_param.hard_swish_threshold =
Expand Down