matt/ollama37
1
0
Fork 0
mirror of https://github.com/dogkeeper886/ollama37.git synced 2025-12-10 15:57:04 +00:00
Code Issues Packages Projects Releases Wiki Activity

relay load model errors to the client (#3065)

Browse Source
This commit is contained in:
Bruce MacDonald
2024-03-11 16:48:27 -04:00
committed by GitHub
parent 6d3adfbea2
commit b80661e8c7
3 changed files with 51 additions and 11 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
llm/ext_server/ext_server.cpp
View File

@@ -114,16 +114,12 @@ void llama_server_init(ext_server_params *sparams, ext_server_resp_t *err) {
llama_backend_init();
llama_numa_init(params.numa);
// load the model
if (!llama->load_model(params)) {
// TODO - consider modifying the logging logic or patching load_model so
// we can capture more detailed error messages and pass them back to the
// caller for better UX
err->id = -1;
snprintf(err->msg, err->msg_len, "error loading model %s",
params.model.c_str());
return;
}
if (!llama->load_model(params)) {
// an error occured that was not thrown
err->id = -1;
snprintf(err->msg, err->msg_len, "error loading model %s", params.model.c_str());
return;
}
llama->initialize();
} catch (std::exception &e) {