furyhawk/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2026-07-21 10:15:53 +00:00
Code Issues Packages Projects Releases Wiki Activity

quant : allow using manual tensor types with --pure (#25716)

Browse Source
This commit is contained in:
Georgi Gerganov
2026-07-16 08:30:20 +03:00
committed by GitHub
parent 3dafb585f8
commit 56d6e9dde2
1 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/llama-quant.cpp
+2 -2
View File
@@ -673,7 +673,7 @@ static ggml_type llama_tensor_get_type(quantize_state_impl & qs, const llama_mod
ggml_type new_type = default_type;
// get more optimal quantization type based on the tensor shape, layer, etc.
if (!params->pure && ggml_is_quantized(default_type)) {
if (ggml_is_quantized(default_type)) {
// if the user provided tensor types - use those
bool manual = false;
if (!qs.tensor_type_patterns.empty()) {
@@ -692,7 +692,7 @@ static ggml_type llama_tensor_get_type(quantize_state_impl & qs, const llama_mod
}
// if not manual - use the standard logic for choosing the quantization type based on the selected mixture
if (!manual) {
if (!manual && !params->pure) {
new_type = llama_tensor_get_type_impl(qs, new_type, tensor, params->ftype, tm.category);
}