@@ -787,10 +787,7 @@ ov::SoPtr<ov::ICompiledModel> ov::CoreImpl::compile_model(const std::string& mod
787
787
788
788
auto cacheManager = coreConfig.get_cache_config_for_device (plugin, parsed._config )._cacheManager ;
789
789
790
- if (plugin.get_name ().find (" LLAMA_CPP" ) != std::string::npos) {
791
- compiled_model = plugin.compile_model (model_path, parsed._config );
792
- }
793
- else if (cacheManager && device_supports_model_caching (plugin) && !is_proxy_device (plugin)) {
790
+ if (cacheManager && device_supports_model_caching (plugin) && !is_proxy_device (plugin)) {
794
791
// Skip caching for proxy plugin. HW plugin will load network from the cache
795
792
CacheContent cacheContent{cacheManager, model_path};
796
793
cacheContent.blobId = ov::ModelCache::compute_hash (model_path, create_compile_config (plugin, parsed._config ));
@@ -800,13 +797,8 @@ ov::SoPtr<ov::ICompiledModel> ov::CoreImpl::compile_model(const std::string& mod
800
797
auto model = read_model (model_path, std::string{});
801
798
return compile_model_and_cache (plugin, model, parsed._config , {}, cacheContent);
802
799
});
803
- } else if (cacheManager) {
804
- // this code path is enabled for AUTO / MULTI / BATCH / PROXY devices which don't support
805
- // import / export explicitly, but can redirect this functionality to actual HW plugin
806
- compiled_model = plugin.compile_model (model_path, parsed._config );
807
800
} else {
808
- auto model = read_model (model_path, std::string ());
809
- compiled_model = plugin.compile_model (model, parsed._config );
801
+ compiled_model = plugin.compile_model (model_path, parsed._config );
810
802
}
811
803
return compiled_model;
812
804
}
0 commit comments