vrr/koboldcpp
2
0
Fork 0
mirror of https://github.com/LostRuins/koboldcpp.git synced 2025-09-16 03:49:42 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

add e5m2 support for use in Kobo, also made a separate contribution PR https://github.com/leejet/stable-diffusion.cpp/pull/460

Browse source
This commit is contained in:
Concedo 2024-11-13 11:35:49 +08:00
parent 3813f6c517
commit dd95f88c19
2 changed files with 67 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

5
otherarch/sdcpp/model.h
View file

@ -35,6 +35,7 @@ struct TensorStorage {
ggml_type type = GGML_TYPE_F32;
bool is_bf16 = false;
bool is_f8_e4m3 = false;
bool is_f8_e5m2 = false;
int64_t ne[SD_MAX_DIMS] = {1, 1, 1, 1, 1};
int n_dims = 0;
@ -64,7 +65,7 @@ struct TensorStorage {
}
int64_t nbytes_to_read() const {
if (is_bf16 || is_f8_e4m3) {
if (is_bf16 || is_f8_e4m3 || is_f8_e5m2) {
return nbytes() / 2;
} else {
return nbytes();
@ -114,6 +115,8 @@ struct TensorStorage {
type_name = "bf16";
} else if (is_f8_e4m3) {
type_name = "f8_e4m3";
} else if (is_f8_e5m2) {
type_name = "f8_e5m2";
}
ss << name << " | " << type_name << " | ";
ss << n_dims << " [";