From 202eab04d3f5a06304b9fd43b4b6b079d3f76dfa Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?M=2E=20Yusuf=20Sar=C4=B1g=C3=B6z?= <yusufsarigoz@gmail.com>
Date: Sat, 12 Aug 2023 16:39:05 +0300
Subject: [PATCH] gguf : quantization is working

---
 examples/gguf/gguf.cpp | 2 +-
 gguf-llama.cpp         | 7 ++-----
 gguf-util.h            | 4 ++++
 3 files changed, 7 insertions(+), 6 deletions(-)

diff --git a/examples/gguf/gguf.cpp b/examples/gguf/gguf.cpp
index 08f2b6322..6f454a204 100644
--- a/examples/gguf/gguf.cpp
+++ b/examples/gguf/gguf.cpp
@@ -421,7 +421,7 @@ int main(int argc, char ** argv) {
         GGML_ASSERT(gguf_ex_write(fname) && "failed to write gguf file");
     } else if (mode == "r") {
         GGML_ASSERT(gguf_ex_read_0(fname) && "failed to read gguf file");
-        //GGML_ASSERT(gguf_ex_read_1(fname) && "failed to read gguf file");
+        GGML_ASSERT(gguf_ex_read_1(fname) && "failed to read gguf file");
         GGML_ASSERT(gguf_ex_read_2(fname) && "failed to read gguf file");
     } else if (mode == "q") {
         llama_model_quantize_params params = llama_model_quantize_default_params();
diff --git a/gguf-llama.cpp b/gguf-llama.cpp
index ea721a0c7..700d6009b 100644
--- a/gguf-llama.cpp
+++ b/gguf-llama.cpp
@@ -752,7 +752,6 @@ struct gguf_file_saver {
         file.seek(info_offset, SEEK_SET);
         GGML_ASSERT(info_offset == file.tell());
         total_written += file.write_str(tensor.name);
-printf("total_written = %zu, name = %s\n", total_written, tensor.name.c_str());
 
         int32_t n_dims = tensor.ne.size();
         total_written += file.write_i32(n_dims);
@@ -765,8 +764,7 @@ printf("total_written = %zu, name = %s\n", total_written, tensor.name.c_str());
         info_offset   += total_written;
 
         file.seek(0, SEEK_END);
-        printf("total_written = %zu\n", total_written);
-
+        
         return total_written;
     }
 
@@ -936,8 +934,7 @@ struct llama_model_loader {
         } else {
             gguf_file & file = file_loader->file;
             file.seek(lt.file_off, SEEK_SET);
-            // TODO
-            //file.read_raw(lt.data, lt.size);
+            file.read_raw(lt.data, lt.size);
         }
 
         if (0) {
diff --git a/gguf-util.h b/gguf-util.h
index ed7d53f69..6395cf304 100644
--- a/gguf-util.h
+++ b/gguf-util.h
@@ -131,6 +131,10 @@ struct gguf_file {
         fwrite(data, size, 1, fp);
     }
 
+    void read_raw(void * data, size_t size) {
+        fread(data, size, 1, fp);
+    }
+
     template<typename T>
     void write_val(const std::string & key, enum gguf_type type, const T & val) {
         write_str(key);