happyz
/
llama.cpp
mirror of https://github.com/ggerganov/llama.cpp.git
1
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

correct function names

This commit is contained in:
ddh0 2026-02-13 21:17:53 -06:00
parent bddc67547f
commit 7b127e126a
1 changed files with 5 additions and 3 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
src/llama-quant.cpp
View File

@ -437,13 +437,15 @@ static ggml_type llama_tensor_get_type_impl(quantize_state_impl & qs, ggml_type
return new_type;
}
// determine the ggml_type that this tensor should be quantized to
// determine the ggml_type that this tensor should be quantized to.
//
// `qs` statistics will only be updated if the `update_stats` parameter is true.
static ggml_type llama_tensor_get_type(
quantize_state_impl & qs,
const llama_model_quantize_params * params,
const ggml_tensor * tensor,
ggml_type default_type,
bool update_stats // we only update qs if this flag is true
bool update_stats
) {
ggml_type new_type = default_type;
// get more optimal quantization type based on the tensor shape, layer, etc.
@ -973,7 +975,7 @@ static void llama_model_quantize_impl(const std::string & fname_inp, const std::
// if so, what will be the target type?
if (do_quantize) {
new_type = tensor_get_target_type(qs, params, tensor, default_type, true);
new_type = llama_tensor_get_type(qs, params, tensor, default_type, true);
// If we've decided to quantize to the same type the tensor is already
// in then there's nothing to do.
do_quantize = tensor->type != new_type;