Results for “metro”

Configure fault tolerance, straggler detection, preemption, in-process restart, and re-run state machine for Megatron Bridge training jobs.

NVIDIA's runtime safety framework for LLM applications. Features jailbreak detection, input/output validation, fact-checking, hallucination detection, PII filtering, toxicity detection. Uses Colang 2.0 DSL for programmable rails. Production-ready, runs on T4 GPU.

NVIDIA's runtime safety framework for LLM applications. Features jailbreak detection, input/output validation, fact-checking, hallucination detection, PII filtering, toxicity detection. Uses Colang 2.0 DSL for programmable rails. Production-ready, runs on T4 GPU.

Run Megatron-LM (MLM) and Megatron Bridge training with mock or real data, covering correlation testing, available recipes, and multi-GPU examples.

Train, evaluate, export, and run inference for monocular depth estimation models using Metric Depth Anything v2 or Relative Depth Anything architectures via the TAO toolkit.

Keep long-running agent work within 5-hour and weekly usage limits by checking usage between waves, pausing near the cap, and resuming only when the window is clear.

Validate and configure packed sequences and long-context training in Megatron-Bridge, distinguishing offline packed SFT for LLMs from in-batch packing for VLMs with correct context parallelism constraints.

Optimizes MoE expert-parallel communication overlap in Megatron Bridge, covering dispatch/combine overlap, flex dispatcher backends, and expert wgrad scheduling.

Runs a three-phase AOI training pipeline: AutoML HPO baseline, DEFT iterative data improvement, and AutoML refinement on the augmented dataset.

How Hermes talks to Gas Town from outside the Mayor session — listing beads, slinging work to rigs, sending mail, escalating, all via the `gt` MCP server (Ophis Cobra→MCP bridge). Use whenever the user asks Hermes to dispatch agent work, check what the swarm is doing, or coordinate across rigs.

Train Mixture of Experts (MoE) models using DeepSpeed or HuggingFace, covering architectures, routing, load balancing, and expert parallelism.

Drive mex (`mex-agent`), persistent project memory and code graphs for AI coding agents. One command scaffolds a living wiki, builds a deterministic code graph, and installs a project anchor file (CLAUDE.md, root AGENTS.md, .cursorrules, .windsurfrules, copilot-instructions.md, or .opencode/opencode.json) that your agent auto-loads as a standing rule document. Use when the user wants to `mex setup` a new project, build a symbol-grounded wiki, keep knowledge connected to implementation, route relevant context to agents, or run drift detection (`mex check`, `mex sync`). Triggers on: "mex setup", "project memory", "code graphs", "codebase documentation", "drift detection", "agent memory", "structured scaffolds", "architectural context", "living wiki", "project anchor file".

Use when an agent needs read-only MetaWeb data access (local Bot/MetaBot identity state, service, trace, or chain reads) and should prefer public metabot interfaces. Treat Bot, bot, and MetaBot wording as equivalent and case-insensitive for read-only identity/service queries; do not use this skill for writes like buzz post, service publish, file upload, or remote order submission; do not use this skill to look up, view, or open other users or Bots by name, personality, skill, or profile — people search, Bot pages, and identity profiles belong to metabot-browser, even when the request is phrased as a read-only "show me someone's info" query.

Rate Limiting — Agent Arena

This skill should be used when the user asks about "key startup metrics", "SaaS metrics", "CAC and LTV", "unit economics", "burn multiple", "rule of 40", "marketplace metrics", or requests guidance on tracking and optimizing business performance metrics.