highperfocused/ollama
1
0
Fork 0
mirror of https://github.com/ollama/ollama.git synced 2025-11-11 14:27:36 +01:00
Code Issues Packages Projects Releases Wiki Activity

api: structured outputs - chat endpoint (#7900)

Browse Source 
Adds structured outputs to chat endpoint
---------

Co-authored-by: Michael Yang <mxyng@pm.me>
Co-authored-by: Hieu Nguyen <hieunguyen1053@outlook.com>
This commit is contained in:
Parth Sareen
2024-12-04 16:31:19 -08:00
committed by GitHub
parent eb8366d658
commit 630e7dc6ff
10 changed files with 180 additions and 25 deletions
Download Patch File Download Diff File Expand all files Collapse all files

29
llama/sampling_ext.cpp vendored
View File

@@ -1,11 +1,13 @@
// TODO: this is a temporary wrapper to allow calling C++ code from CGo
#include "sampling.h"
#include "sampling_ext.h"
#include "json-schema-to-grammar.h"
struct gpt_sampler *gpt_sampler_cinit(
const struct llama_model *model, struct gpt_sampler_cparams *params)
{
try {
try
{
gpt_sampler_params sparams;
sparams.top_k = params->top_k;
sparams.top_p = params->top_p;
@@ -24,7 +26,9 @@ struct gpt_sampler *gpt_sampler_cinit(
sparams.seed = params->seed;
sparams.grammar = params->grammar;
return gpt_sampler_init(model, sparams);
} catch (const std::exception & err) {
}
catch (const std::exception &err)
{
return nullptr;
}
}
@@ -54,3 +58,24 @@ void gpt_sampler_caccept(
{
gpt_sampler_accept(sampler, id, apply_grammar);
}
int schema_to_grammar(const char *json_schema, char *grammar, size_t max_len)
{
try
{
nlohmann::json schema = nlohmann::json::parse(json_schema);
std::string grammar_str = json_schema_to_grammar(schema);
size_t len = grammar_str.length();
if (len >= max_len)
{
len = max_len - 1;
}
strncpy(grammar, grammar_str.c_str(), len);
return len;
}
catch (const std::exception &e)
{
strncpy(grammar, "", max_len - 1);
return 0;
}
}