CS348Project/llama.cpp
5
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-11-03 09:22:01 +00:00
Code Issues Packages Projects Releases Wiki Activity

gguf : start implementing gguf_file_saver (WIP)

Browse Source
This commit is contained in:
M. Yusuf Sarıgöz
2023-08-11 11:29:50 +03:00
parent a356b0e228
commit b2440f1943
1 changed files with 12 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

15
gguf-llama.cpp
View File

@@ -647,9 +647,14 @@ struct gguf_file_saver {
}
void write_hparams(enum llama_ftype new_ftype) {
const llama_hparams & hparams = any_file_loader->hparams;
GGML_UNUSED(hparams);
GGML_UNUSED(new_ftype);
const int32_t n_kv = gguf_get_n_kv(any_file_loader->gguf_ctx);
for (int i = 0; i < n_kv; ++i) {
const char * key = gguf_get_key(any_file_loader->gguf_ctx, i);
if (strcmp(key, "general.quantization_version") == 0) {
file.write_val<uint32_t>("general.quantization_version", GGUF_TYPE_UINT32, new_ftype);
}
}
}
void write_vocab() {
@@ -658,6 +663,10 @@ struct gguf_file_saver {
}
void write_tensor(llama_load_tensor & tensor, enum ggml_type new_type, const void * new_data, size_t new_size) {
GGML_UNUSED(tensor);
GGML_UNUSED(new_data);
GGML_UNUSED(new_size);
switch (new_type) {
case GGML_TYPE_F32:
case GGML_TYPE_F16: