lora : add support for non-llama models (#3333)

* lora : add support for non-llama models

ggml-ci

* avoid leaking ggml_context on failure
cleanup

ggml-ci

* lora : allow 1d tensors

* lora : include embd and output layers in size calculation

* fix style

1 files changed, 1 insertions, 0 deletions

diff --git a/llama.h b/llama.h
index 45a65cac..15ab4f80 100644
--- a/llama.h
+++ b/llama.h
@@ -39,6 +39,7 @@
 
 #define LLAMA_MAX_RNG_STATE (64*1024)
 
+#define LLAMA_FILE_MAGIC_GGLA 0x67676c61u // 'ggla'
 #define LLAMA_FILE_MAGIC_GGSN 0x6767736eu // 'ggsn'
 
 #define LLAMA_SESSION_MAGIC   LLAMA_FILE_MAGIC_GGSN