[ Frontend ] Multiprocessing for OpenAI Server with zeromq (#6883)

Signed-off-by: Joe Runde <Joseph.Runde@ibm.com>
Co-authored-by: Joe Runde <Joseph.Runde@ibm.com>
Co-authored-by: Joe Runde <joe@joerun.de>
Co-authored-by: Nick Hill <nickhill@us.ibm.com>
Co-authored-by: Simon Mo <simon.mo@hey.com>
This commit is contained in:
Robert Shaw
2024-08-02 18:27:28 -07:00
committed by GitHub
co-authored by Joe Runde Joe Runde Nick Hill Simon Mo
parent 708989341e
commit ed812a73fa
20 changed files with 1567 additions and 101 deletions
+7 -2
View File
@@ -131,9 +131,14 @@ def make_arg_parser(parser: FlexibleArgumentParser) -> FlexibleArgumentParser:
parser.add_argument(
"--return-tokens-as-token-ids",
action="store_true",
help="When --max-logprobs is specified, represents single tokens as"
"strings of the form 'token_id:{token_id}' so that tokens that"
help="When --max-logprobs is specified, represents single tokens as "
"strings of the form 'token_id:{token_id}' so that tokens that "
"are not JSON-encodable can be identified.")
parser.add_argument(
"--disable-frontend-multiprocessing",
action="store_true",
help="If specified, will run the OpenAI frontend server in the same "
"process as the model serving engine.")
parser = AsyncEngineArgs.add_cli_args(parser)