Erik Scholz
d8c65b4436
add inplace conversion support for f8_e4m3 ( #359 )
...
in the same way it is done for bf16
like how bf16 converts losslessly to fp32,
f8_e4m3 converts losslessly to fp16
2024-08-24 12:23:09 +08:00
Erik Scholz and leejet
f2e4d9793b
fix: avoid some memory leaks ( #136 )
...
---------
Co-authored-by: leejet <leejet714@gmail.com >
2024-01-01 23:27:29 +08:00
Erik Scholz
4a5e7b58e2
fix: never use a log message as a format string ( #135 )
2024-01-01 20:43:47 +08:00
Erik Scholz
f469b835a3
fix: reading memory of stack allocated object past its scope ( #91 )
2023-11-27 21:37:12 +08:00
Erik Scholz
c874063408
fix: support bf16 lora weights ( #82 )
2023-11-20 22:34:17 +08:00
Erik Scholz and leejet
844351c417
feat: cmake improvements and simple ci ( #9 )
...
* move main and stb-libs to subfolder
* cmake : general additions
* ci : add simple building
---------
Co-authored-by: leejet <31925346+leejet@users.noreply.github.com >
2023-08-17 21:09:57 +08:00
Erik Scholz