Public/llama.cpp
1
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2026-09-09 22:39:04 +02:00
Code Issues Packages Projects Releases Wiki Activity

Enable backend sampling with token speculation

Browse Source
This commit is contained in:
Gaurav Garg
2026-07-01 18:21:47 +05:30
parent 360e1349f0
commit 167d5e141b
18 changed files with 764 additions and 319 deletions
Download Patch File Download Diff File Expand all files Collapse all files
tests/test-arg-parser.cpp
+10
View File
@@ -2,7 +2,9 @@
#include "common.h"
#include "download.h"
#include "llama.h"
#include "speculative.h"
#include <limits>
#include <string>
#include <vector>
#include <sstream>
@@ -14,6 +16,14 @@
static void test(void) {
common_params params;
assert(common_speculative_n_outputs_max(16, 2, 3) == 8);
assert(common_speculative_n_outputs_max(16, 2, -1) == 2);
assert(common_speculative_n_outputs_max(4, 2, 3) == 4);
assert(common_speculative_n_outputs_max(
std::numeric_limits<int32_t>::max(),
std::numeric_limits<int32_t>::max(),
std::numeric_limits<int32_t>::max()) == std::numeric_limits<int32_t>::max());
printf("test-arg-parser: make sure there is no duplicated arguments in any examples\n\n");
for (int ex = 0; ex < LLAMA_EXAMPLE_COUNT; ex++) {
try {