EngineX-Mthreads/enginex-mthreads-llama.cpp
4
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

Revert "sync from b7516"

Browse Source 
This reverts commit 6ee41dd9e3.
This commit is contained in:
xiezhongtao
2026-01-16 11:34:48 +08:00
parent 6ee41dd9e3
commit 620f04050b
380 changed files with 38842 additions and 18471 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
docs/backend/hexagon/developer.md
View File

@@ -49,7 +49,7 @@ Each Hexagon device behaves like a GPU from the offload and model splitting pers
Here is an example of running GPT-OSS-20B model on a newer Snapdragon device with 16GB of DDR.
```
M=gpt-oss-20b-Q4_0.gguf NDEV=4 D=HTP0,HTP1,HTP2,HTP3 P=surfing.txt scripts/snapdragon/adb/run-cli.sh -no-cnv -f surfing.txt -n 32
M=gpt-oss-20b-Q4_0.gguf NDEV=4 D=HTP0,HTP1,HTP2,HTP3 P=surfing.txt scripts/snapdragon/adb/run-completion.sh -f surfing.txt -n 32
...
LD_LIBRARY_PATH=/data/local/tmp/llama.cpp/lib
ADSP_LIBRARY_PATH=/data/local/tmp/llama.cpp/lib