Distribute and run LLMs with a single file.
C++ cross-platformggufllama-cpplocal-ailocal-inferencelocal-llmopen-source-aisingle-file-executablespeech-to-text
diff --git a/examples/common.cpp b/examples/common.cpp
--- a/whisper.cpp/examples/common.cpp
+++ b/whisper.cpp/examples/common.cpp
@@ -2,6 +2,7 @@
#include "common.h"
+#include <algorithm>
#include <cmath>
#include <codecvt>
#include <cstring>
@@ -418,9 +419,9 @@ gpt_vocab::id gpt_sample_top_k_top_p(
}
// find the top K tokens
- std::partial_sort(
+ std::sort(
logits_id.begin(),
- logits_id.begin() + top_k, logits_id.end(),
+ logits_id.end(),
[](const std::pair<double, gpt_vocab::id> & a, const std::pair<double, gpt_vocab::id> & b) {
return a.first > b.first;
});
@@ -533,9 +534,9 @@ gpt_vocab::id gpt_sample_top_k_top_p_repeat(
}
// find the top K tokens
- std::partial_sort(
+ std::sort(
logits_id.begin(),
- logits_id.begin() + top_k, logits_id.end(),
+ logits_id.end(),
[](const std::pair<double, gpt_vocab::id> & a, const std::pair<double, gpt_vocab::id> & b) {
return a.first > b.first;
});