A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 16, 23:15 UTC
[Bugfix] Fix FlexibleArgumentParser replaces _ with - for actual args (#5795)
Chang Su committed
Jun 24, 2024 at 16:01 UTC
ba991d5c84adbc0685075af88333c688ddb06011
1 file changed
+7
-1
vllm/utils.py
+7
-1
@@ -822,7 +822,13 @@ class FlexibleArgumentParser(argparse.ArgumentParser):
822
processed_args = []
823
for arg in args:
824
if arg.startswith('--'):
825
- processed_args.append('--' + arg[len('--'):].replace('_', '-'))
825
+ if '=' in arg:
826
+ key, value = arg.split('=', 1)
827
+ key = '--' + key[len('--'):].replace('_', '-')
828
+ processed_args.append(f'{key}={value}')
829
+ else:
830
+ processed_args.append('--' +
831
+ arg[len('--'):].replace('_', '-'))
832
else:
833
processed_args.append(arg)
834