Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Add profile subcommand to run perf analyzer #13

Merged
merged 3 commits into from
Dec 15, 2023
Merged

Add profile subcommand to run perf analyzer #13

Show file tree
Hide file tree
Changes from 1 commit
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
e415cd3
Add profile subcommand to run perf analyzer
matthewkotila Dec 15, 2023
b5829d0
Fix precommit hooks
nv-hwoo Dec 15, 2023
2e205cb
Revert back the input tensor names
nv-hwoo Dec 15, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
16 changes: 15 additions & 1 deletion src/triton_cli/parser.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,6 +9,7 @@
from triton_cli.metrics import MetricsClient
from triton_cli.repository import ModelRepository
from triton_cli.server.server_factory import TritonServerFactory
from triton_cli.profiler import Profiler

logger = logging.getLogger("triton")

Expand Down Expand Up @@ -84,6 +85,17 @@ def handle_model(args: argparse.Namespace):

if args.subcommand == "infer":
client.infer(args.model, args.data, args.prompt)
elif args.subcommand == "profile":
# TODO: run PA LLM benchmark script
print("pull engine()")
print("run_server()")
print("profile()")
Profiler.profile(
model=args.model,
batch_size=64,
url="localhost:8001",
input_length=2048,
)
elif args.subcommand == "config":
config = client.get_model_config(args.model)
if config:
Expand Down Expand Up @@ -160,11 +172,13 @@ def parse_args_model(subcommands):
default=None,
help="Text input to LLM-like models. Required for inference on LLMs, optional otherwise.",
)
profile = model_commands.add_parser("profile", help="Run Perf Analyzer")
profile.add_argument("-m", "--model", type=str, required=True, help="Model name")
config = model_commands.add_parser("config", help="Get config for model")
config.add_argument("-m", "--model", type=str, required=True, help="Model name")
metrics = model_commands.add_parser("metrics", help="Get metrics for model")
metrics.add_argument("-m", "--model", type=str, required=True, help="Model name")
add_client_args([infer, config, metrics])
add_client_args([infer, profile, config, metrics])
return model


Expand Down
Loading
Loading