37 lines
1.0 KiB
YAML
37 lines
1.0 KiB
YAML
# SPDX-FileCopyrightText: Copyright (c) 2025-2026 NVIDIA CORPORATION & AFFILIATES. All rights reserved.
|
|
# SPDX-License-Identifier: Apache-2.0
|
|
|
|
apiVersion: nvidia.com/v1alpha1
|
|
kind: DynamoGraphDeployment
|
|
metadata:
|
|
name: mocker-agg
|
|
spec:
|
|
services:
|
|
Frontend:
|
|
componentType: frontend
|
|
replicas: 1
|
|
extraPodSpec:
|
|
mainContainer:
|
|
image: my-registry/mocker-runtime:my-tag
|
|
decode:
|
|
envFromSecret: hf-token-secret
|
|
componentType: worker
|
|
subComponentType: decode
|
|
replicas: 1
|
|
extraPodSpec:
|
|
mainContainer:
|
|
image: my-registry/mocker-runtime:my-tag
|
|
workingDir: /workspace
|
|
command:
|
|
- python3
|
|
- -m
|
|
- dynamo.mocker
|
|
args:
|
|
- --model-path
|
|
- nvidia/Llama-3.1-8B-Instruct-FP8
|
|
- --model-name
|
|
- nvidia/Llama-3.1-8B-Instruct-FP8
|
|
- --speedup-ratio
|
|
- "1.0"
|
|
- --planner-profile-data
|
|
- /workspace/tests/planner/profiling_results/H200_TP1P_TP1D |