runtime-llamacpp-v0.1.3

PaddlePaddle/Paddleruntime-llamacpp-v0.1.3Jun 28, 2026by LauraGPT

AI Summary

This release is a portable rebuild of the runtime binaries designed to fix stability issues. It disables aggressive CPU optimizations found in the runner environment to ensure the binaries work on a wider range of generic CPUs without crashing.

Key Highlights

Rebuilds v0.1.2 with conservative CPU flags
Disables ggml native CPU tuning and high-ISA flags
Resolves crashes caused by Illegal instruction or SIGILL errors

New Features

Conservative build for broader CPU compatibility
SIGILL crash fix

Full Release Notes

Portable rebuild of the FunASR llama.cpp / GGUF runtime binaries.

This release rebuilds the v0.1.2 runtime with conservative CPU flags for the generic x64 assets. In particular, it disables ggml native CPU tuning and high-ISA AVX/AVX2/AVX512/VNNI/FMA/F16C/BMI2 flags so the prebuilt binaries do not depend on the GitHub Actions runner CPU instruction set.

Use this version if v0.1.2 crashes with `Illegal instruction`, `SIGILL`, or Windows error `0xC000001D`.