「serving」の検索結果

Itôで予約済みのGPUを使ったモデル配信の相談に対し、現状の未対応範囲と将来必要な確認事項を整理するスキル。現在は配信を開始せず、制約を案内します。

• Itô GPU予約後の配信対応を確認したいとき
• OpenAI互換の接続先の制約を知りたいとき
• 将来のモデル配信設定を確認したいとき

[omh] Self-hosted LLM serving on GPUs: choose the serving engine and quantization from decision tables, prepare deployment as an idempotent runbook with observed-only verification, and measure the endpoint with the standard TTFT/TPOT/goodput protocol. Use when the user says: inference-serving, inference serving, serve this model, serve the model, model serving, serving endpoint, vllm, llama.cpp.

日本語の概要は準備中です。原文の説明を表示しています。

Use when querying or calling pay-per-token foundation models, or when creating, updating, listing, or querying Unity Gateway (also called Unity AI Gateway) services, including model provider services, MCP services, system.ai services, and three-part Unity Catalog service names (catalog.schema.service). Also use for managing spend budgets and usage alerts on Unity Gateway, and for migrating from legacy AI Gateway. Also use when a workload with Model Serving endpoint permissions (for example CAN_QUERY) gets PERMISSION_DENIED querying a three-part model name, since serving-endpoint access does not grant access to a Unity Gateway service. Not for configuring legacy AI Gateway on Model Serving endpoints; use databricks-model-serving for that.

日本語の概要は準備中です。原文の説明を表示しています。

Serves LLMs with high throughput using vLLM's PagedAttention and continuous batching. Use when deploying production LLM APIs, optimizing inference latency/throughput, or serving models with limited GPU memory. Supports OpenAI-compatible endpoints, quantization (GPTQ/AWQ/FP8), and tensor parallelism.

日本語の概要は準備中です。原文の説明を表示しています。

Serves LLMs with high throughput using vLLM's PagedAttention and continuous batching. Use when deploying production LLM APIs, optimizing inference latency/throughput, or serving models with limited GPU memory. Supports OpenAI-compatible endpoints, quantization (GPTQ/AWQ/FP8), and tensor parallelism.

日本語の概要は準備中です。原文の説明を表示しています。

Deploy ML models on Kubernetes with KServe (formerly KFServing) and NVIDIA Triton Inference Server. Includes canary deployments, autoscaling, model versioning, A/B testing, and GPU resource management for production model serving.

日本語の概要は準備中です。原文の説明を表示しています。

LLM and ML model deployment for inference. Use when serving models in production, building AI APIs, or optimizing inference. Covers vLLM (LLM serving), TensorRT-LLM (GPU optimization), Ollama (local), BentoML (ML deployment), Triton (multi-model), LangChain (orchestration), LlamaIndex (RAG), and streaming patterns.

日本語の概要は準備中です。原文の説明を表示しています。

Use when running throughput/latency (serving speed) benchmarks against an already-running model server. Uses sgl_jax.bench_serving to sweep batch-size/concurrency and measure throughput and latency, producing summary.csv and per-point bs_<N>/result.jsonl. Speed only — accuracy/eval belongs to model-accuracy-benchmark. Triggers on throughput, latency, bench_serving, batch-size sweep, tok/s, TTFT, TPOT — not accuracy/eval.

日本語の概要は準備中です。原文の説明を表示しています。

Databricks Model Serving endpoint lifecycle and ops. Use when asked to: CRUD serving endpoints (CLI or MLflow Deployments client); configure traffic routing for A/B / canary deploys and zero-downtime version swaps; retrieve OpenAPI schemas; inspect logs, metrics, or permissions; manage legacy AI Gateway rate limits (not Unity Gateway); discover Foundation Model API endpoints at runtime; integrate endpoints into Databricks Apps; or stream from off-platform clients (Vercel AI SDK v6, standalone Node.js). NOT for: Unity Gateway CRUD and management (databricks-unity-gateway), training, MLflow autologging, UC registration, custom PyFunc/ResponsesAgent authoring (databricks-ml-training); Knowledge Assistants/Supervisor Agents (databricks-agent-bricks); MLflow evaluation (databricks-mlflow-evaluation).

日本語の概要は準備中です。原文の説明を表示しています。

Design privacy-preserving analytics systems using differential privacy, k-anonymity, l-diversity, and t-closeness. Covers privacy budget allocation with epsilon tracking, references Google DP library, OpenDP, and Apple PPML. Includes Python differential privacy implementation for GDPR-compliant statistical analysis.

日本語の概要は準備中です。原文の説明を表示しています。

Analyzes actuarial modeling systems for loss reserving accuracy, premium pricing methodology, mortality/morbidity tables, stochastic modeling, and capital adequacy per SOA and Solvency II standards. USE THIS SKILL WHEN: - You need to review or audit actuarial models (reserving, pricing, capital) - Someone asks about loss triangle analysis or reserve adequacy - You are evaluating IBNR calculations, chain ladder methods, or Bornhuetter-Ferguson - A project involves insurance pricing, GLM rating models, or ratemaking - You need to assess Solvency II SCR calculations or RBC compliance - Someone mentions actuarial opinions, ASOP compliance, or SOA standards - You are reviewing stochastic models, ESG configurations, or DFA frameworks - A codebase uses actuarial libraries (chainladder, lifetables, ChainLadder R package) TRIGGER PHRASES: "actuarial", "loss reserving", "IBNR", "chain ladder", "premium pricing", "mortality table", "Solvency II", "capital adequacy", "ratemaking", "GLM pricing", "risk-based capital", "reserve analysis", "actuarial opinion"

日本語の概要は準備中です。原文の説明を表示しています。

Audit current model deployment topology — serving config, latency profile, version inventory. Use when asked "what models are deployed", "audit our serving setup", or "inventory our model versions".

日本語の概要は準備中です。原文の説明を表示しています。

Design and configure model serving infrastructure — endpoint scaling, batching, GPU allocation. Use when asked to "serve this model", "design an inference endpoint", or "size GPU allocation for serving".

日本語の概要は準備中です。原文の説明を表示しています。

Practical naked-eye and binocular sky observing. Covers dark adaptation, limiting magnitude, constellation recognition, star hopping, the messier and Caldwell catalogs accessible without a telescope, the Moon, planets, meteor showers, aurorae, and the ethics and habits of observing under light-polluted and dark skies. Use when teaching someone to find their way around the sky, planning a naked-eye session, or choosing a first binocular tour.

日本語の概要は準備中です。原文の説明を表示しています。

Deploy machine learning models to production serving infrastructure using MLflow, BentoML, or Seldon Core with REST/gRPC endpoints, implement autoscaling, monitoring, and A/B testing capabilities for high-performance model inference at scale. Use when deploying trained models for real-time inference, setting up REST or gRPC prediction APIs, implementing autoscaling for variable load, running A/B tests between model versions, or migrating from batch to real-time inference.

日本語の概要は準備中です。原文の説明を表示しています。

Build a feature store using Feast for centralized feature management, configure offline and online stores for batch and real-time serving, define feature views with transformations, and implement point-in-time correct joins for ML pipelines. Use when managing features for multiple ML models, ensuring training-serving consistency, serving low-latency features for real-time inference, reusing feature definitions across projects, or building a feature catalog for discovery and governance.

日本語の概要は準備中です。原文の説明を表示しています。

Design model serving infrastructure for AI. TRIGGERS - Use when user needs help with ai-model-serving related tasks.

日本語の概要は準備中です。原文の説明を表示しています。

Pick the right serving container for a SageMaker model deployment and find its current image URI. Use this skill whenever about to deploy a model to a SageMaker endpoint and an image URI needs to be chosen — including when the user says "deploy this LLM", "host this HuggingFace model", "serve this fine-tuned model", "deploy this embedding model", "host a reranker", "serve a sentence-transformers model", or when about to hardcode any container URI in deployment code. HuggingFace-curated Deep Learning Containers are ALWAYS preferred: HuggingFace vLLM (LLMs and generative rerankers), HuggingFace vLLM-Omni (multimodal), TEI (embeddings/cross-encoder rerankers), HF Inference Toolkit (other transformers). Generic images (AWS vLLM, DJL-LMI, SGLang) are used only when no HuggingFace image is compatible — never merely because they carry a newer version. Never hardcode a container URI from memory and never default to TGI. Prevents stale-image failures and wrong-region URIs.

日本語の概要は準備中です。原文の説明を表示しています。

Design model serving infrastructure for AI. TRIGGERS - Use when user needs help with ai-model-serving related tasks.

日本語の概要は準備中です。原文の説明を表示しています。

OpenClawのテストやプラグイン検証の遅さ・メモリ使用量を計測し、読み込みや起動処理の原因を調査。テストの検証範囲を保ちながら改善し、前後の数値を比較します。

• 遅いテストや重い読み込みの原因調査
• プラグイン検証の時間とメモリ削減
• テスト中のメモリ増加を調べたいとき

コードの動作を保ちながら、重複削除やモジュール分割で構造を整えるスキル。変更前と各段階のテスト、計画の承認、レビュー、コミット前の確認まで進行を管理します。

• 動作を変えずにモジュールを分割したいとき
• 重複コードや不要コードの削除
• 入れ子の削減と名前の整理

vLLM: high-throughput LLM serving, OpenAI API, quantization.

日本語の概要は準備中です。原文の説明を表示しています。

NVIDIA GPU上で大規模言語モデルの応答生成を高速化し、APIとして提供する設定を支援します。量子化や一括処理、複数GPUへの分散も扱います。

• 言語モデルをチャットAPIで提供したいとき
• 多数のプロンプトを一括処理したいとき
• 量子化でメモリ使用量を抑えたいとき

Modal is a serverless cloud platform for running Python on demand, including on-demand GPUs. Use when deploying or serving AI/ML models, running GPU-accelerated workloads (training, fine-tuning, inference), serving web endpoints, scheduling batch jobs, or scaling Python code to cloud containers with the Modal SDK.

日本語の概要は準備中です。原文の説明を表示しています。