backend : add eval callback (#4935)

* backend : add eval callback

ggml-ci

* backend : group nodes in a single compute when user don't need them

* backend : clean-up the implementation

ggml-ci

* simple : do not perform tensor data copy if not needed

* simple : fix

* simple : no need for ggml_is_contiguous + fix bool parse

* llama : fix callback placement in llama_context_params

* backend : avoid double-ask callback calls

* simple : restore examples, imatrix will serve as a demo

1 files changed, 4 insertions, 0 deletions

diff --git a/llama.h b/llama.h
index a570b0d6..e268d7a1 100644
--- a/llama.h
+++ b/llama.h
@@ -2,6 +2,7 @@
 #define LLAMA_H
 
 #include "ggml.h"
+#include "ggml-backend.h"
 #ifdef GGML_USE_CUBLAS
 #include "ggml-cuda.h"
 #define LLAMA_MAX_DEVICES GGML_CUDA_MAX_DEVICES
@@ -231,6 +232,9 @@ extern "C" {
         float    yarn_beta_slow;   // YaRN high correction dim
         uint32_t yarn_orig_ctx;    // YaRN original context size
 
+        ggml_backend_sched_eval_callback cb_eval;
+        void * cb_eval_user_data;
+
         enum ggml_type type_k; // data type for K cache
         enum ggml_type type_v; // data type for V cache