Mirrors/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggerganov/llama.cpp.git synced 2024-12-27 22:59:24 +01:00
Code Issues Packages Projects Releases Wiki Activity

ggml: add support for float16 input tensors in pooling operations (ggml/895)

Browse Source 
* Add support for float16 tensors in 1d pooling operations

* Add support for float16 input tensors in 2d pooling operations

* code cleanup

remove unnecessary casting during srow ptr initialization

---------

Co-authored-by: vanaka11 <vanaka1189@gmail.com>
This commit is contained in:
Ivan Filipov 2024-07-22 14:32:02 +03:00 committed by Georgi Gerganov
parent 203b7f1531
commit 9f77d899b7
1 changed files with 12 additions and 12 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

20
ggml/src/ggml.c
View File

@ -14746,7 +14746,7 @@ static void ggml_compute_forward_pool_1d_sk_p0(
const struct ggml_tensor * src = dst->src[0]; const struct ggml_tensor * src = dst->src[0];
assert(src->type == GGML_TYPE_F32); assert(src->type == GGML_TYPE_F32 || src->type == GGML_TYPE_F16);
if (params->ith != 0) { if (params->ith != 0) {
return; return;
@ -14759,10 +14759,8 @@ static void ggml_compute_forward_pool_1d_sk_p0(
const int64_t rs = dst->ne[0]; const int64_t rs = dst->ne[0];
while (cdata < data_end) { while (cdata < data_end) {
const float * const srow = (const float *)cdata; const void * srow = (const void *)cdata;
int j = 0; int j = 0;
for (int64_t i = 0; i < rs; ++i) { for (int64_t i = 0; i < rs; ++i) {
switch (op) { switch (op) {
case GGML_OP_POOL_AVG: drow[i] = 0; break; case GGML_OP_POOL_AVG: drow[i] = 0; break;
@ -14770,9 +14768,10 @@ static void ggml_compute_forward_pool_1d_sk_p0(
case GGML_OP_POOL_COUNT: GGML_ABORT("fatal error"); case GGML_OP_POOL_COUNT: GGML_ABORT("fatal error");
} }
for (int ki = 0; ki < k; ++ki) { for (int ki = 0; ki < k; ++ki) {
const float srow_j = (src->type == GGML_TYPE_F32) ? ((const float*)srow)[j] : GGML_FP16_TO_FP32(((const ggml_fp16_t*)srow)[j]);
switch (op) { switch (op) {
case GGML_OP_POOL_AVG: drow[i] += srow[j]; break; case GGML_OP_POOL_AVG: drow[i] += srow_j; break;
case GGML_OP_POOL_MAX: if (srow[j] > drow[i]) drow[i] = srow[j]; break; case GGML_OP_POOL_MAX: if (srow_j > drow[i]) drow[i] = srow_j; break;
case GGML_OP_POOL_COUNT: GGML_ABORT("fatal error"); case GGML_OP_POOL_COUNT: GGML_ABORT("fatal error");
} }
++j; ++j;
@ -14814,7 +14813,7 @@ static void ggml_compute_forward_pool_2d(
const struct ggml_tensor * src = dst->src[0]; const struct ggml_tensor * src = dst->src[0];
GGML_ASSERT(src->type == GGML_TYPE_F32); assert(src->type == GGML_TYPE_F32 || src->type == GGML_TYPE_F16);
if (params->ith != 0) { if (params->ith != 0) {
return; return;
@ -14857,13 +14856,14 @@ static void ggml_compute_forward_pool_2d(
for (int ky = 0; ky < k1; ++ky) { for (int ky = 0; ky < k1; ++ky) {
if (iy + ky < 0 || iy + ky >= src->ne[1]) continue; if (iy + ky < 0 || iy + ky >= src->ne[1]) continue;
const float * const srow = (const float *)(cdata + src->nb[1] * (iy + ky)); const void * srow = (const void *)(cdata + src->nb[1] * (iy + ky));
for (int kx = 0; kx < k0; ++kx) { for (int kx = 0; kx < k0; ++kx) {
int j = ix + kx; int j = ix + kx;
if (j < 0 || j >= src->ne[0]) continue; if (j < 0 || j >= src->ne[0]) continue;
const float srow_j = (src->type == GGML_TYPE_F32) ? ((const float*)srow)[j] : GGML_FP16_TO_FP32(((const ggml_fp16_t*)srow)[j]);
switch (op) { switch (op) {
case GGML_OP_POOL_AVG: *out += srow[j]; break; case GGML_OP_POOL_AVG: *out += srow_j; break;
case GGML_OP_POOL_MAX: if (srow[j] > *out) *out = srow[j]; break; case GGML_OP_POOL_MAX: if (srow_j > *out) *out = srow_j; break;
case GGML_OP_POOL_COUNT: GGML_ABORT("fatal error"); case GGML_OP_POOL_COUNT: GGML_ABORT("fatal error");
} }
} }