“dpo”的搜索结果

Implement and debug Google OAuth 2.0 refresh tokens bound to a client key. Covers key generation, signed DPoP proofs that demonstrate key possession, and nonce challenge retries in WebCrypto and JavaScript environments.

中文简介正在准备中。当前显示英语简介。

• Bind Google OAuth refresh tokens to a client key
• Debug DPoP proof signatures and nonce retries
• Plan a backend for a browser application using DPoP

Choose a model training method and configuration from paired response preferences or individual positive and negative ratings. Covers DPO, ORPO, KTO, and SimPO, including pair construction, repeated training rounds, and troubleshooting learning settings.

中文简介正在准备中。当前显示英语简介。

• Choose a training method for paired response preferences
• Select a method for individual positive and negative ratings
• Review learning settings when DPO training loses general capability

Manages Agent Platform endpoints, the stable URLs used to serve online model predictions. It supports listing, inspecting, creating, updating, and deleting endpoints, plus troubleshooting permission errors, quota limits, and resources that are busy.

中文简介正在准备中。当前显示英语简介。

• Listing endpoints and checking their configuration
• Creating an endpoint before model deployment
• Updating endpoint display names or labels

Deploy SageMaker endpoints, the services that answer model requests, with automatic capacity scaling, CloudWatch alarms, and resource tags. Supports real-time and asynchronous inference, scaling down to zero instances, deployment checks, and resource cleanup.

中文简介正在准备中。当前显示英语简介。

• Deploying a real-time model endpoint
• Scaling a sparse workload down to zero instances
• Running long inference jobs asynchronously

Explore and diagnose a PostHog endpoint's execution logs — error messages, failed runs, cache misses, slow runs, or unexpected row counts during endpoint invocations. Use when the user says "my endpoint is failing", "show me the logs for endpoint X", "what error did endpoint Y produce", "why did endpoint Z return no rows", "is this endpoint hitting cache", or "check the last N runs". Focused on a single named endpoint's runtime log entries, not project-wide auditing or query performance profiling.

中文简介正在准备中。当前显示原文说明。

Turn agent failure records into training examples by locating the first incorrect decision and pairing preferred actions with rejected ones. Assess whether training is appropriate, then check both targeted improvements and preservation of correct behavior.

中文简介正在准备中。当前显示英语简介。

• Diagnosing the first incorrect decision in an agent failure
• Building preference pairs from failure records
• Choosing between system fixes and model training

Deploy osquery on Windows, macOS, and Linux to inspect processes, listening ports, software, and system settings using SQL, a database query language. Configure scheduled checks and centralized management to support threat hunting and compliance monitoring.

中文简介正在准备中。当前显示英语简介。

• Deploying osquery across Windows, macOS, and Linux
• Investigating suspicious processes and network connections with SQL
• Periodically checking software inventories and system settings

Diagnose failed model requests from OpenShell sandboxes to hosted APIs or local servers such as Ollama and vLLM. Check provider attachment, access rules, credentials, network layout, and client settings to identify the narrowest configuration change needed.

中文简介正在准备中。当前显示英语简介。

• Diagnose failed requests to hosted model APIs
• Troubleshoot connections to Ollama or LM Studio
• Investigate credential and endpoint mismatch errors

Create a SageMaker endpoint (real-time, real-time scale-to-zero, or async) with autoscaling, CloudWatch alarms, and tagging enabled by default. Use this skill whenever about to create a SageMaker endpoint, write deployment code that calls `create_endpoint`, or finalize a deployment after the image URI and IAM role are known. Provides deploy.py for real-time endpoints, deploy_ic.py for real-time endpoints that scale to zero instances via inference components, and deploy_async.py for async endpoints (also scale-to-zero). This is the last step in the SageMaker deployment workflow. Never generate a bare `create_endpoint` call without these defaults — endpoints without autoscaling or alarms are demos, not deployments.

中文简介正在准备中。当前显示原文说明。

Deploy CrowdStrike Falcon sensors across Windows, macOS, and Linux endpoints. Configure threat prevention and response policies, verify installation and detection, troubleshoot sensor issues, and connect events to SIEM platforms, which collect security logs for analysis.

中文简介正在准备中。当前显示英语简介。

• Deploying Falcon sensors across enterprise endpoints
• Configuring prevention and response policies by endpoint group
• Checking sensor connectivity and test detections

Configure advanced Microsoft Defender protections on Windows, including rules that block risky behavior, protected folders, network filtering, and exploit defenses. Use audit modes to assess application impact, then deploy policies and monitor detections across an organization.

中文简介正在准备中。当前显示英语简介。

• Strengthening Windows endpoint protection
• Assessing ASR rules in audit mode
• Configuring ransomware protection for folders

Guide response to suspected or confirmed malware infections across enterprise endpoints. Identify the malware and entry route, assess affected systems, coordinate containment and removal, and document recovery checks, infection timelines, and observable signs of compromise.

中文简介正在准备中。当前显示英语简介。

• Investigating malware alerts from endpoint security tools
• Assessing infection spread across multiple endpoints
• Planning containment, removal, and endpoint recovery

Deploy Model Garden models to Agent Platform endpoints, the addresses used for model requests. Check supported hardware and estimated hourly costs, inspect deployment progress, test predictions, and remove deployed resources after use.

中文简介正在准备中。当前显示英语简介。

• Choose a deployable model and supported hardware
• Estimate hourly deployment costs before approval
• Check deployment progress and test predictions

Guide language model training with SimPO using pairs of preferred and rejected answers. Covers configuration, learning rate tuning, and memory troubleshooting for preference alignment, which adjusts model responses toward human preferences without a separate reference model.

中文简介正在准备中。当前显示英语简介。

• Training on preferred and rejected answer pairs
• Exploring an alternative to DPO without a reference model
• Tuning preference training for math and code tasks

Guide large language model training with OpenRLHF using human preference feedback. Covers training method selection, distributed GPU setup, reward model training, and memory tuning, with examples for PPO, GRPO, and DPO.

中文简介正在准备中。当前显示英语简介。

• Choosing a training method for human preference alignment
• Setting up distributed training across multiple GPUs
• Training a reward model before PPO

Document and type a Sentry API endpoint. Write or fix @extend_schema decorators, specify response TypedDicts, type request parameters, correct type drift between the declared schema and the runtime response, and validate the generated spec. Use when asked to "document an endpoint", "add OpenAPI docs", "add/fix @extend_schema", "type an endpoint response", "fix the response type", "fix type drift", "reuse a response type", "split an overloaded endpoint", "specify the response schema", "add a TypedDict response", "migrate a legacy api-docs path", "fix a parameter type", or "make an endpoint public" / "promote an endpoint" (promotion is one section here).

中文简介正在准备中。当前显示原文说明。

arkcli 推理接入点管理与显式 raw CRUD 创建能力。**正向触发：用户明确要求脚本化 / CI / 无护栏 / 原始 raw CRUD 创建 Endpoint 时，必须使用本 skill 的 `arkcli infer endpoint create` 路径，不能切到 `arkcli-deploy`**；模型缺失或只有品牌/家族名时仍留在本 skill，先执行实时有界候选与 0/1/N 澄清，模型未唯一确定前禁止 preview 或创建。本 skill 也负责对**已有** Endpoint 做获取、列表、启动、停止、删除、更新，以及在 NotFound 后用当前完整列表 + 历史 usage 判断 ID 是否不完整或仅有历史记录。普通的『创建/新建/部署一个 endpoint』仍走带产品护栏的 arkcli-deploy（`+deploy`），只有用户明确选择上述原始路径才例外。优先使用产品命令 `arkcli infer endpoint ...`，而不是 Raw API。TTS/ASR/语音模型连 raw create 也不要引导，只能转 models search 说明 arkcli 不支持 Endpoint 创建。

中文简介正在准备中。当前显示原文说明。

Deploys and configures osquery for real-time endpoint monitoring using SQL-based queries to inspect running processes, open ports, installed software, and system configuration. Use when building visibility into endpoint state, threat hunting across fleet, or implementing compliance monitoring. Activates for requests involving osquery deployment, endpoint visibility, fleet management, or SQL-based endpoint querying.

中文简介正在准备中。当前显示原文说明。

Deploys and configures osquery for real-time endpoint monitoring using SQL-based queries to inspect running processes, open ports, installed software, and system configuration. Use when building visibility into endpoint state, threat hunting across fleet, or implementing compliance monitoring. Activates for requests involving osquery deployment, endpoint visibility, fleet management, or SQL-based endpoint querying.

中文简介正在准备中。当前显示原文说明。

Document and type a Sentry API endpoint. Write or fix @extend_schema decorators, specify response TypedDicts, type request parameters, correct type drift between the declared schema and the runtime response, and validate the generated spec. Use when asked to "document an endpoint", "add OpenAPI docs", "add/fix @extend_schema", "type an endpoint response", "fix the response type", "fix type drift", "reuse a response type", "split an overloaded endpoint", "specify the response schema", "add a TypedDict response", "migrate a legacy api-docs path", "fix a parameter type", or "make an endpoint public" / "promote an endpoint" (promotion is one section here).

中文简介正在准备中。当前显示原文说明。

Deploys and configures osquery for real-time endpoint monitoring using SQL-based queries to inspect running processes, open ports, installed software, and system configuration. Use when building visibility into endpoint state, threat hunting across fleet, or implementing compliance monitoring. Activates for requests involving osquery deployment, endpoint visibility, fleet management, or SQL-based endpoint querying.

中文简介正在准备中。当前显示原文说明。

Guide language model training with TRL using example answers, preferred response pairs, or reward scores. The skill covers data preparation, method selection, reward model training, evaluation, and adjustments for memory limits or unstable training.

中文简介正在准备中。当前显示英语简介。

• Train a language model to follow instructions using example answers
• Train on preferred and rejected responses with DPO
• Build a reward model and run an RLHF training pipeline

Examine Windows endpoint logs for attempts to avoid security detection, including log clearing, altered file timestamps, injected code, and disabled protection. Correlate events to investigate suspicious behavior and develop detection rules with context for legitimate activity.

中文简介正在准备中。当前显示英语简介。

• Investigating cleared logs and altered file timestamps
• Checking suspected attempts to disable security tools
• Finding abuse of legitimate Windows utilities

Plan and configure large language model training with OpenRLHF to better match human preferences. Covers training method selection, distributed GPU setup, inference acceleration with vLLM, and troubleshooting memory or training stability issues.

中文简介正在准备中。当前显示英语简介。

• Choosing a method for training models on human preferences
• Setting up PPO training with a reward model
• Configuring GRPO training without a critic model