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

Disable TRT-LLM echo by default #194

Open
wants to merge 2 commits into
base: main
Choose a base branch
from
Open

Disable TRT-LLM echo by default #194

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
7da4f0b
Disable TRT-LLM echo by default
dyastremsky Nov 23, 2024
a2f62cd
Update docs
dyastremsky Nov 23, 2024
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
5 changes: 2 additions & 3 deletions genai-perf/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -388,9 +388,8 @@ a request in order. Random means that assignment is uniformly random

##### `--backend {tensorrtllm,vllm}`

When using the "triton" service-kind, this is the backend of the model. For the
TRT-LLM backend, you currently must set `exclude_input_in_output` to true in the
model config to not echo the input tokens in the output. (default: tensorrtllm)
When using the "triton" service-kind, this is the backend of the model.
(default: tensorrtllm)

##### `--endpoint <str>`

Expand Down
1 change: 1 addition & 0 deletions genai-perf/genai_perf/inputs/converters/tensorrtllm_converter.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -60,6 +60,7 @@ def convert(
"model": model_name,
"text_input": [text],
"max_tokens": [DEFAULT_TENSORRTLLM_MAX_TOKENS], # default
"exclude_input_in_output": [True], # default
}
self._add_request_params(payload, config)
request_body["data"].append(payload)
Expand Down
5 changes: 1 addition & 4 deletions genai-perf/genai_perf/parser.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -644,10 +644,7 @@ def _add_endpoint_args(parser):
default="tensorrtllm",
required=False,
help=f'When using the "triton" service-kind, '
"this is the backend of the model. "
"For the TENSORRT-LLM backend, you currently must set "
"'exclude_input_in_output' to true in the model config to "
"not echo the input tokens in the output.",
"this is the backend of the model. ",
)

endpoint_group.add_argument(
Expand Down
4 changes: 4 additions & 0 deletions genai-perf/tests/test_triton_tensorrtllm_converter.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -75,11 +75,13 @@ def test_convert_default(self):
"model": "test_model",
"text_input": ["text input one"],
"max_tokens": [DEFAULT_TENSORRTLLM_MAX_TOKENS],
"exclude_input_in_output": [True],
},
{
"model": "test_model",
"text_input": ["text input two"],
"max_tokens": [DEFAULT_TENSORRTLLM_MAX_TOKENS],
"exclude_input_in_output": [True],
},
]
}
Expand Down Expand Up @@ -116,6 +118,7 @@ def test_convert_with_request_parameters(self):
"max_tokens": [1234],
"stream": [True],
"additional_key": ["additional_value"],
"exclude_input_in_output": [True],
},
{
"model": "test_model",
Expand All @@ -124,6 +127,7 @@ def test_convert_with_request_parameters(self):
"max_tokens": [1234],
"stream": [True],
"additional_key": ["additional_value"],
"exclude_input_in_output": [True],
},
]
}
Expand Down
5 changes: 2 additions & 3 deletions templates/genai-perf-templates/README_template
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -386,9 +386,8 @@ a request in order. Random means that assignment is uniformly random

##### `--backend {tensorrtllm,vllm}`

When using the "triton" service-kind, this is the backend of the model. For the
TRT-LLM backend, you currently must set `exclude_input_in_output` to true in the
model config to not echo the input tokens in the output. (default: tensorrtllm)
When using the "triton" service-kind, this is the backend of the model.
(default: tensorrtllm)

##### `--endpoint <str>`

Expand Down
Loading