pingu_98/llama_cpp_for_radxa_dragon_wing_q6a
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions
Modifiers to get Llama.cpp working using the built in NPU on the Radxa Dragon Wing Q6A SBC (Qualcomm QCS6490 cpu). Hacked together with Claude Code and Deepseek V4 Flash. It works, but the overall performance for TG is poor, ingestion is super fast - but
1 commit 1 branch 2 tags 334 MiB
C++ 45.3%
JavaScript 20.2%
C 10.4%
Python 6.2%
Cuda 4.7%
Other 13%
Find a file
Georgi Gerganov 26c0846629
Initial release
2023-03-10 20:56:40 +02:00
.gitignore Initial release 2023-03-10 20:56:40 +02:00
convert-pth-to-ggml.py Initial release 2023-03-10 20:56:40 +02:00
ggml.c Initial release 2023-03-10 20:56:40 +02:00
ggml.h Initial release 2023-03-10 20:56:40 +02:00
main.cpp Initial release 2023-03-10 20:56:40 +02:00
Makefile Initial release 2023-03-10 20:56:40 +02:00
quantize.cpp Initial release 2023-03-10 20:56:40 +02:00
utils.cpp Initial release 2023-03-10 20:56:40 +02:00
utils.h Initial release 2023-03-10 20:56:40 +02:00