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

Add F16/BF16 type

This commit is contained in:
Ed Addario 2025-08-20 13:23:11 +01:00
parent 5cd69a6809
commit 69586e212e
No known key found for this signature in database
GPG Key ID: E7875815A3230993
1 changed files with 4 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
tools/quantize/quantize.cpp
View File

@ -132,7 +132,7 @@ static void usage(const char * executable) {
printf(" Advanced option to selectively quantize tensors. May be specified multiple times.\n"); printf(" Advanced option to selectively quantize tensors. May be specified multiple times.\n");
printf(" --prune-layers L0,L1,L2...comma-separated list of layer numbers to prune from the model\n"); printf(" --prune-layers L0,L1,L2...comma-separated list of layer numbers to prune from the model\n");
printf(" Advanced option to remove all tensors from the given layers\n"); printf(" Advanced option to remove all tensors from the given layers\n");
printf(" --target-bpw: target bits per weight (bpw). Must be a positive number between 0.0 and 8.0 \n"); printf(" --target-bpw: target bits per weight (bpw). Must be a positive number between 0.0 and 16.0 \n");
printf(" --keep-split: will generate quantized model in the same shards as input\n"); printf(" --keep-split: will generate quantized model in the same shards as input\n");
printf(" --override-kv KEY=TYPE:VALUE\n"); printf(" --override-kv KEY=TYPE:VALUE\n");
printf(" Advanced option to override model metadata by key in the quantized model. May be specified multiple times.\n"); printf(" Advanced option to override model metadata by key in the quantized model. May be specified multiple times.\n");
@ -486,13 +486,13 @@ static bool parse_target_bpw(const char * data, float & target_bpw) {
try { try {
target_bpw = std::stof(data); target_bpw = std::stof(data);
if (target_bpw < 0.0f || target_bpw > 8.0f) { if (target_bpw < 0.0f || target_bpw > 16.0f) {
printf("\n%s: target bits per weight (bpw) must be a positive number between 0.0 and 8.0\n\n", __func__); printf("\n%s: target bits per weight (bpw) must be a positive number between 0.0 and 16.0\n\n", __func__);
return false; return false;
} }
} }
catch (const std::exception & e) { catch (const std::exception & e) {
printf("\n%s: '%s' is not valid. Target bits per weight (bpw) must be a positive number between 0.0 and 8.0\n\n", __func__, data); printf("\n%s: '%s' is not valid. Target bits per weight (bpw) must be a positive number between 0.0 and 16.0\n\n", __func__, data);
return false; return false;
} }