From 69586e212e76849fcdff17e68e8023b91025b415 Mon Sep 17 00:00:00 2001 From: Ed Addario Date: Wed, 20 Aug 2025 13:23:11 +0100 Subject: [PATCH] Add F16/BF16 type --- tools/quantize/quantize.cpp | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-) diff --git a/tools/quantize/quantize.cpp b/tools/quantize/quantize.cpp index 86a96cdfcc..b907008cb4 100644 --- a/tools/quantize/quantize.cpp +++ b/tools/quantize/quantize.cpp @@ -132,7 +132,7 @@ static void usage(const char * executable) { 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(" 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(" --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"); @@ -486,13 +486,13 @@ static bool parse_target_bpw(const char * data, float & target_bpw) { try { target_bpw = std::stof(data); - if (target_bpw < 0.0f || target_bpw > 8.0f) { - printf("\n%s: target bits per weight (bpw) must be a positive number between 0.0 and 8.0\n\n", __func__); + 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 16.0\n\n", __func__); return false; } } 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; }