upstream-archive/stable-diffusion.cpp
1
0
Fork 0
mirror of https://github.com/leejet/stable-diffusion.cpp.git synced 2026-10-03 03:17:51 -05:00
Code Issues Packages Projects Releases Wiki Activity

feat: load scaled FP8 weights without upfront conversion (#1913)

Browse Source
This commit is contained in:
leejet
2026-08-27 20:50:47 +08:00
committed by GitHub
parent 50d6405683
commit be0e34480d
9 changed files with 73 additions and 101 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/model_io/safetensors_io.cpp
+4 -6
View File
@@ -87,9 +87,9 @@ static ggml_type safetensors_dtype_to_ggml_type(const std::string& dtype) {
} else if (dtype == "F64") {
ttype = GGML_TYPE_F32;
} else if (dtype == "F8_E4M3") {
ttype = GGML_TYPE_F16;
ttype = GGML_TYPE_F8_E4M3;
} else if (dtype == "F8_E5M2") {
ttype = GGML_TYPE_F16;
ttype = GGML_TYPE_F8_E5M2;
} else if (dtype == "I32") {
ttype = GGML_TYPE_I32;
} else if (dtype == "I64") {
@@ -328,12 +328,10 @@ bool read_safetensors_file(const std::string& file_path,
bool tensor_size_ok;
if (dtype == "F8_E4M3") {
tensor_storage.is_f8_e4m3 = true;
// f8 -> f16
tensor_size_ok = (tensor_storage.nbytes() == tensor_data_size * 2);
tensor_size_ok = (tensor_storage.nbytes() == tensor_data_size);
} else if (dtype == "F8_E5M2") {
tensor_storage.is_f8_e5m2 = true;
// f8 -> f16
tensor_size_ok = (tensor_storage.nbytes() == tensor_data_size * 2);
tensor_size_ok = (tensor_storage.nbytes() == tensor_data_size);
} else if (dtype == "F64") {
tensor_storage.is_f64 = true;
// f64 -> f32
src/model_io/tensor_storage.h
+1 -3
View File
@@ -54,9 +54,7 @@ struct TensorStorage {
}
int64_t nbytes_to_read() const {
if (is_f8_e4m3 || is_f8_e5m2) {
return nbytes() / 2;
} else if (is_f64 || is_i64) {
if (is_f64 || is_i64) {
return nbytes() * 2;
} else {
return nbytes();