From 4201fa5cb854118b4f172a758e4bcd421866f3bf Mon Sep 17 00:00:00 2001
From: Georgi Gerganov <ggerganov@gmail.com>
Date: Tue, 9 May 2023 18:19:13 +0300
Subject: [PATCH] llama : produce error upon loading old model files

---
 llama.cpp | 15 +++++++++++++--
 llama.h   |  2 +-
 2 files changed, 14 insertions(+), 3 deletions(-)

diff --git a/llama.cpp b/llama.cpp
index 367522ac9..334d4e1bc 100644
--- a/llama.cpp
+++ b/llama.cpp
@@ -402,6 +402,7 @@ enum llama_file_version {
     LLAMA_FILE_VERSION_GGML,
     LLAMA_FILE_VERSION_GGMF_V1, // added version field and scores in vocab
     LLAMA_FILE_VERSION_GGJT_V1, // added padding
+    LLAMA_FILE_VERSION_GGJT_V2, // changed quantization format
 };
 
 struct llama_file_loader {
@@ -432,6 +433,8 @@ struct llama_file_loader {
             file_version = LLAMA_FILE_VERSION_GGMF_V1;
         } else if (magic == 'ggjt' && version == 1) {
             file_version = LLAMA_FILE_VERSION_GGJT_V1;
+        } else if (magic == 'ggjt' && version == 2) {
+            file_version = LLAMA_FILE_VERSION_GGJT_V2;
         } else {
             throw format("unknown (magic, version) combination: %08x, %08x; is this really a GGML file?",
                          magic, version);
@@ -837,8 +840,8 @@ static const char *llama_file_version_name(llama_file_version version) {
     switch (version) {
         case LLAMA_FILE_VERSION_GGML: return "'ggml' (old version with low tokenizer quality and no mmap support)";
         case LLAMA_FILE_VERSION_GGMF_V1: return "ggmf v1 (old version with no mmap support)";
-        case LLAMA_FILE_VERSION_GGJT_V1: return "ggjt v1 (latest)";
-        default: LLAMA_ASSERT(false);
+        case LLAMA_FILE_VERSION_GGJT_V1: return "ggjt v1 (pre #1305)";
+        case LLAMA_FILE_VERSION_GGJT_V2: return "ggjt v2 (latest)";
     }
 }
 
@@ -915,6 +918,14 @@ static void llama_model_load_internal(
         fprintf(stderr, "%s: model size = %s\n",  __func__, llama_model_type_name(model.type));
     }
 
+    if (file_version != LLAMA_FILE_VERSION_GGJT_V2) {
+        if (hparams.ftype != LLAMA_FTYPE_ALL_F32     &&
+            hparams.ftype != LLAMA_FTYPE_MOSTLY_F16  &&
+            hparams.ftype != LLAMA_FTYPE_MOSTLY_Q8_0) {
+            throw format("this format is no longer supported (see https://github.com/ggerganov/llama.cpp/pull/1305)");
+        }
+    }
+
     if (vocab_only) {
         return;
     }
diff --git a/llama.h b/llama.h
index fea5ffeff..1a65cd589 100644
--- a/llama.h
+++ b/llama.h
@@ -19,7 +19,7 @@
 #    define LLAMA_API
 #endif
 
-#define LLAMA_FILE_VERSION           1
+#define LLAMA_FILE_VERSION           2
 #define LLAMA_FILE_MAGIC             'ggjt'
 #define LLAMA_FILE_MAGIC_UNVERSIONED 'ggml'
 #define LLAMA_SESSION_MAGIC          'ggsn'