From ec7506d680cf6b40d7a5a140dc25bab1805e37d0 Mon Sep 17 00:00:00 2001
From: City <125218114+city96@users.noreply.github.com>
Date: Thu, 27 Nov 2025 22:05:57 +0100
Subject: [PATCH] Flux2 specific filter

---
 tools/lcpp.patch | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)

diff --git a/tools/lcpp.patch b/tools/lcpp.patch
index 3e449ac..975ee26 100644
--- a/tools/lcpp.patch
+++ b/tools/lcpp.patch
@@ -39,7 +39,7 @@ index b16c462f..6d1568f1 100644
      const int idx = gguf_find_tensor(ctx, name);
      if (idx < 0) {
 diff --git a/src/llama.cpp b/src/llama.cpp
-index 24e1f1f0..8a1e9ef8 100644
+index 24e1f1f0..78b74deb 100644
 --- a/src/llama.cpp
 +++ b/src/llama.cpp
 @@ -205,6 +205,18 @@ enum llm_arch {
@@ -354,7 +354,7 @@ index 24e1f1f0..8a1e9ef8 100644
      }
  
      // Set split info if needed
-@@ -18647,6 +18908,124 @@ static void llama_model_quantize_internal(const std::string & fname_inp, const s
+@@ -18647,6 +18908,127 @@ static void llama_model_quantize_internal(const std::string & fname_inp, const s
          // do not quantize relative position bias (T5)
          quantize &= name.find("attn_rel_b.weight") == std::string::npos;
  
@@ -368,6 +368,9 @@ index 24e1f1f0..8a1e9ef8 100644
 +            quantize &= name.find("vector_in.") == std::string::npos;
 +            quantize &= name.find("guidance_in.") == std::string::npos;
 +            quantize &= name.find("final_layer.") == std::string::npos;
++            // flux2
++            quantize &= name.find("single_stream_modulation.") == std::string::npos;
++            quantize &= name.find("double_stream_modulation_") == std::string::npos;
 +        }
 +        if (model.arch == LLM_ARCH_SD1 || model.arch == LLM_ARCH_SDXL) {
 +            image_model = true;
@@ -479,7 +482,7 @@ index 24e1f1f0..8a1e9ef8 100644
          enum ggml_type new_type;
          void * new_data;
          size_t new_size;
-@@ -18655,6 +19034,9 @@ static void llama_model_quantize_internal(const std::string & fname_inp, const s
+@@ -18655,6 +19037,9 @@ static void llama_model_quantize_internal(const std::string & fname_inp, const s
              new_type = default_type;
  
              // get more optimal quantization type based on the tensor shape, layer, etc.
@@ -489,7 +492,7 @@ index 24e1f1f0..8a1e9ef8 100644
              if (!params->pure && ggml_is_quantized(default_type)) {
                  new_type = llama_tensor_get_type(qs, new_type, tensor, ftype);
              }
-@@ -18664,6 +19046,7 @@ static void llama_model_quantize_internal(const std::string & fname_inp, const s
+@@ -18664,6 +19049,7 @@ static void llama_model_quantize_internal(const std::string & fname_inp, const s
              if (params->output_tensor_type < GGML_TYPE_COUNT && strcmp(tensor->name, "output.weight") == 0) {
                  new_type = params->output_tensor_type;
              }