Skip to content
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

ggml : fix quants nans when all the group weights are very close to zero #7313

Merged
merged 3 commits into from
May 18, 2024
Merged

ggml : fix quants nans when all the group weights are very close to zero #7313

Show file tree
Hide file tree
Changes from 2 commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
6fa6a9a
ggml : fix quants nans when all the group weights are very close to zero
slaren May 15, 2024
f59edee
increase eps to 1e-7
slaren May 15, 2024
f07e570
use higher eps only for the quants that need it
slaren May 17, 2024
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
22 changes: 12 additions & 10 deletions ggml-quants.c
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -14,6 +14,8 @@
#include <stdlib.h> // for qsort
#include <stdio.h> // for GGML_ASSERT

#define GROUP_MAX_EPS 1e-7f

#if defined(_MSC_VER)
// disable "possible loss of data" to avoid warnings for hundreds of casts
// we should just be careful :)
Expand Down Expand Up @@ -1109,7 +1111,7 @@ static float make_qx_quants(int n, int nmax, const float * restrict x, int8_t *
float ax = fabsf(x[i]);
if (ax > amax) { amax = ax; max = x[i]; }
}
if (amax < 1e-30f) { // all zero
if (amax < GROUP_MAX_EPS) { // all zero
for (int i = 0; i < n; ++i) {
L[i] = 0;
}
Expand Down Expand Up @@ -1177,7 +1179,7 @@ static float make_q3_quants(int n, int nmax, const float * restrict x, int8_t *
float ax = fabsf(x[i]);
if (ax > amax) { amax = ax; max = x[i]; }
}
if (!amax) { // all zero
if (amax < GROUP_MAX_EPS) { // all zero
for (int i = 0; i < n; ++i) { L[i] = 0; }
return 0.f;
}
Expand Down Expand Up @@ -2653,7 +2655,7 @@ void quantize_row_q6_K_reference(const float * restrict x, block_q6_K * restrict

}

if (!max_abs_scale) {
if (max_abs_scale < GROUP_MAX_EPS) {
memset(&y[i], 0, sizeof(block_q6_K));
y[i].d = GGML_FP32_TO_FP16(0.f);
x += QK_K;
Expand Down Expand Up @@ -2805,7 +2807,7 @@ static void quantize_row_q6_K_impl(const float * restrict x, block_q6_K * restri

}

if (!max_abs_scale) {
if (max_abs_scale < GROUP_MAX_EPS) {
memset(&y[i], 0, sizeof(block_q6_K));
y[i].d = GGML_FP32_TO_FP16(0.f);
x += QK_K;
Expand Down Expand Up @@ -12772,7 +12774,7 @@ static void quantize_row_iq2_xs_impl(const float * restrict x, void * restrict v
}
float max = xval[0];
for (int i = 1; i < 16; ++i) max = MAX(max, xval[i]);
if (!max) {
if (max < GROUP_MAX_EPS) {
scales[ib] = 0;
memset(L, 0, 16);
continue;
Expand Down Expand Up @@ -13213,7 +13215,7 @@ static void quantize_row_iq3_xxs_impl(int grid_size, const float * restrict x, v
}
float max = xval[0];
for (int i = 1; i < 32; ++i) max = MAX(max, xval[i]);
if (!max) {
if (max < GROUP_MAX_EPS) {
scales[ib] = 0;
memset(L, 0, 32);
continue;
Expand Down Expand Up @@ -13753,7 +13755,7 @@ static void quantize_row_iq1_s_impl(const float * restrict x, void * restrict vy
for (int i = 0; i < block_size; ++i) weight[i] = qw[i] * sqrtf(sigma2 + xb[i]*xb[i]);
float max = fabsf(xb[0]);
for (int i = 1; i < block_size; ++i) max = MAX(max, fabsf(xb[i]));
if (!max) {
if (max < GROUP_MAX_EPS) {
scales[ib] = 0;
memset(L, 1, block_size);
continue;
Expand Down Expand Up @@ -13941,7 +13943,7 @@ static void quantize_row_iq1_m_impl(const float * restrict x, void * restrict vy
}
float max = fabsf(xb[0]);
for (int i = 1; i < block_size; ++i) max = MAX(max, fabsf(xb[i]));
if (!max) {
if (max < GROUP_MAX_EPS) {
scales[ib] = 0;
memset(L, 1, block_size);
continue;
Expand Down Expand Up @@ -14205,7 +14207,7 @@ static void quantize_row_iq4_nl_impl(const int super_block_size, const int block
amax = ax; max = xb[j];
}
}
if (!amax) {
if (amax < GROUP_MAX_EPS) {
scales[ib] = 0;
continue;
}
Expand Down Expand Up @@ -14426,7 +14428,7 @@ static void quantize_row_iq2_s_impl(const float * restrict x, void * restrict vy
}
float max = xval[0];
for (int i = 1; i < 16; ++i) max = MAX(max, xval[i]);
if (!max) {
if (max < GROUP_MAX_EPS) {
scales[ib] = 0;
continue;
}
Expand Down
9 changes: 9 additions & 0 deletions tests/test-backend-ops.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -49,6 +49,15 @@ static void init_tensor_uniform(ggml_tensor * tensor, float min = -1.0f, float m
t.join();
}

#if 0
// test quantization with very small values that may result in nan scales due to division by zero
if (ggml_is_quantized(tensor->type)) {
for (int i = 0; i < 256; i++) {
data[i] = 1e-7f;
}
}
#endif

if (tensor->type == GGML_TYPE_F32 || tensor->type == GGML_TYPE_I32) {
ggml_backend_tensor_set(tensor, data.data(), 0, size * sizeof(float));
} else if (ggml_is_quantized(tensor->type) || tensor->type == GGML_TYPE_F16 || tensor->type == GGML_TYPE_BF16) {
Expand Down
Loading