added an option to offload vae decoder to CPU for mem-limited gpus

stable-diffusion.h

@@ -116,7 +116,8 @@ SD_API sd_ctx_t* new_sd_ctx(const char* model_path,
                             enum sd_type_t wtype,
                             enum rng_type_t rng_type,
                             enum schedule_t s,
-                            bool keep_control_net_cpu);
+                            bool keep_control_net_cpu,
+                            bool keep_vae_on_cpu);
 
 SD_API void free_sd_ctx(sd_ctx_t* sd_ctx);