Agent skill
deploying-triton
Deploys and manages NVIDIA Triton Inference Server containers. Automates model repository setup, config generation, and health checks. Use for "triton 서버", "triton 실행", "모델 서빙", "inference server" requests.
Install this agent skill to your Project
npx add-skill https://github.com/jiunbae/agent-skills/tree/main/ml/triton-deploy
SKILL.md
Triton Deployment
NVIDIA Triton Inference Server management.
Quick Start
# Pull Triton image
docker pull nvcr.io/nvidia/tritonserver:24.01-py3
# Run server
docker run --gpus all -p 8000:8000 -p 8001:8001 -p 8002:8002 \
-v $(pwd)/models:/models \
nvcr.io/nvidia/tritonserver:24.01-py3 \
tritonserver --model-repository=/models
Model Repository Structure
models/
└── my_model/
├── config.pbtxt
└── 1/
└── model.onnx
Config Template
# config.pbtxt
name: "my_model"
platform: "onnxruntime_onnx"
max_batch_size: 8
input [
{ name: "input", data_type: TYPE_FP32, dims: [3, 224, 224] }
]
output [
{ name: "output", data_type: TYPE_FP32, dims: [1000] }
]
Health Check
# Ready check
curl localhost:8000/v2/health/ready
# Model status
curl localhost:8000/v2/models/my_model
Inference
# HTTP
curl -X POST localhost:8000/v2/models/my_model/infer \
-H "Content-Type: application/json" \
-d '{"inputs": [{"name": "input", "shape": [1,3,224,224], "datatype": "FP32", "data": [...]}]}'
# gRPC
grpcurl -d '...' localhost:8001 inference.GRPCInferenceService/ModelInfer
Best Practices
- Use dynamic batching for throughput
- Enable model warmup
- Monitor with Prometheus metrics (:8002)
- Use model versioning (1/, 2/, etc.)
Recommended Agent Skills
Expand your agent's capabilities with these related and highly-rated skills.
jiunbae/agent-skills
implementing-in-background
Orchestrates multiple AI agents (Claude, Codex, Gemini) for parallel implementation in the background. Separates independent tasks from planning docs, each agent writes code directly. Context-safe with auto-save. Use for "백그라운드 구현", "bg impl", "병렬 구현", "Codex로 구현", "구현해줘", "코드 작성해줘" requests.
jiunbae/agent-skills
review-fix-loop
Autonomous review-fix cycle that continuously reviews code using background-reviewer, fixes issues, and repeats until all findings are resolved. Use for "리뷰 루프", "자동 개선", "review fix loop", "리뷰 반복", "코드 개선 루프", "keep reviewing" requests.
jiunbae/agent-skills
planning-in-background
Orchestrates multiple AI agents (Claude, Codex, Gemini) for parallel planning in the background with auto-save. Agents continue running even when session hits context limits. Use for "백그라운드 기획", "bg plan", "병렬 기획", "멀티 AI 기획", "기획해줘", "N명이 기획", "계획", "플래닝", "plan", "설계" requests.
jiunbae/agent-skills
background-reviewer
Orchestrates multi-LLM parallel code review using Claude, Codex, and Gemini. Each agent reviews from a different perspective using agent personas (security, architecture, code quality, performance). Supports persona-based review via `agt persona review`. Use for "코드 리뷰", "리뷰해줘", "bg review", "멀티 리뷰", "background review", "페르소나 리뷰" requests.
jiunbae/agent-skills
managing-context
Discovers and loads relevant project context from markdown documentation before each task. Matches context documents based on keywords, file paths, and task types. Use at task start to access project plans, architecture, and implementation status.
jiunbae/agent-skills
indexing-static-context
Provides an index of global static context files in ~/.agents/. Returns appropriate static file paths for natural language queries like "내 정보", "보안 규칙". Use when other skills or agents need to locate reference information.
Didn't find tool you were looking for?