ggml-jni: sync with upstream llama.cpp

zhouwg / kantv

workbench for learing&practising AI tech in real scenario on Android device, powered by GGML(Georgi Gerganov Machine Learning) and NCNN(Tencent NCNN) and FFmpeg

Apache License 2.0

117 stars 19 forks source link

ggml-jni: sync with upstream llama.cpp #114

Closed zhouwg closed 5 months ago

zhouwg commented 5 months ago

validated with 7 case (benchmark transcription, benchmark mulmat, benchmark memcpy, benchmark whisper-full, benchmakr mulmat, benchmark matrix, benchmark llama, English realtime subtilte for English onlineTV, llama inference) on Xiaomi 14

zhouwg commented 5 months ago

looks good to me.

sometimes crash in llama inference. need more check and find out the root cause.