From 182cbe5d646c9a86cdcdf5dc80193553cb98afc1 Mon Sep 17 00:00:00 2001 From: Colin Kealty <3266127+bartowski1182@users.noreply.github.com> Date: Tue, 10 Mar 2026 11:45:43 -0400 Subject: [PATCH] Fix merge conflicts --- src/llama-quant.cpp | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff --git a/src/llama-quant.cpp b/src/llama-quant.cpp index 79c57832bf..82d7234c7c 100644 --- a/src/llama-quant.cpp +++ b/src/llama-quant.cpp @@ -913,7 +913,8 @@ static void llama_model_quantize_impl(const std::string & fname_inp, const std:: imatrix_data = static_cast>*>(params->imatrix); if (imatrix_data) { LLAMA_LOG_INFO("\n%s: have importance matrix data with %d entries\n", - __func__, (int)imatrix_data->size()); qs.has_imatrix = true; + __func__, (int)imatrix_data->size()); + qs.has_imatrix = true; // check imatrix for nans or infs for (const auto & kv : *imatrix_data) { for (float f : kv.second) { @@ -995,7 +996,7 @@ static void llama_model_quantize_impl(const std::string & fname_inp, const std:: }); } - { // Based on old loop + { std::vector tensor_names; tensor_names.reserve(tensors.size()); for (const auto * it : tensors) { @@ -1005,7 +1006,6 @@ static void llama_model_quantize_impl(const std::string & fname_inp, const std:: } int idx = 0; - uint16_t n_split = 1; // Assume split index is continuous @@ -1048,6 +1048,8 @@ static void llama_model_quantize_impl(const std::string & fname_inp, const std:: const struct ggml_tensor * tensor = it->tensor; const std::string name = ggml_get_name(tensor); + metadata[i].category = tensor_get_category(name); + uint16_t i_split = params->keep_split ? it->idx : 0; if (!ctx_outs[i_split]) { ctx_outs[i_split].reset(gguf_init_empty());