vbatts/llama.cpp
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

parse_args improvements

Browse source
This commit is contained in:
Galunid 2023-11-07 23:48:17 +01:00
parent b7148838f5
commit 6a83bce114
1 changed files with 3 additions and 3 deletions
Download patch file Download diff file Expand all files Collapse all files

6
convert-hf-to-gguf.py
View file

@ -837,13 +837,13 @@ def parse_args() -> argparse.Namespace:
help="path to write to; default: based on input",
)
parser.add_argument(
"--outtype", type=str, choices=["f32", "f16"], default="f16", nargs='?',
help="output format - use 0 for float32, 1 for float16",
"--outtype", type=str, choices=["f32", "f16"], default="f16",
help="output format - use f32 for float32, f16 for float16",
)
parser.add_argument("--bigendian", action="store_true", help="model is executed on big endian machine")
parser.add_argument(
"model", type=Path,
help="directory containing model file, or model file itself (*.bin)",
help="directory containing model file",
)
return parser.parse_args()