“memory”的搜索结果

Reference for the Claude API / Anthropic SDK — model ids, pricing, params, streaming, tool use, MCP, agents, caching, token counting, model migration. TRIGGER — read BEFORE opening the target file; don't skip because it "looks like a one-liner" — whenever: the prompt names Claude/Anthropic in any form (Claude, Anthropic, Fable, Opus, Sonnet, Haiku, `anthropic`, `@anthropic-ai`, `claude-*`, `us.anthropic.*`, `[1m]`); the user asks about an LLM (pricing/model choice/limits/caching) — never answer from memory; OR the task is LLM-shaped with provider unstated (agent/MCP/tool-definition/multi-agent/RAG/LLM-judge/computer-use; generate/summarize/extract/classify/rewrite/converse over NL; debugging refusals/cutoffs/streaming/tool-calls/tokens). SKIP only when another provider is being worked on (overrides all triggers): OpenAI/GPT/Gemini/Llama/Mistral/Cohere/Ollama named in the query; OR `grep -rE 'openai|langchain_openai|google.generativeai|genai|mistralai|cohere|ollama'` over the project hits (run this grep FIRST if no provider named — don't Read the file).

中文简介正在准备中。当前显示原文说明。

Half-Quadratic Quantization for LLMs without calibration data. Use when quantizing models to 4/3/2-bit precision without needing calibration datasets, for fast quantization workflows, or when deploying with vLLM or HuggingFace Transformers.

中文简介正在准备中。当前显示原文说明。

Language-independent tokenizer treating text as raw Unicode. Supports BPE and Unigram algorithms. Fast (50k sentences/sec), lightweight (6MB memory), deterministic vocabulary. Used by T5, ALBERT, XLNet, mBART. Train on raw text without pre-tokenization. Use when you need multilingual support, CJK languages, or reproducible tokenization.

中文简介正在准备中。当前显示原文说明。

Runs LLM inference on CPU, Apple Silicon, and consumer GPUs without NVIDIA hardware. Use for edge deployment, M1/M2/M3 Macs, AMD/Intel GPUs, or when CUDA is unavailable. Supports GGUF quantization (1.5-8 bit) for reduced memory and 4-10× speedup vs PyTorch on CPU.

中文简介正在准备中。当前显示原文说明。

Use this skill for processing and analyzing large tabular datasets (billions of rows) that exceed available RAM. Vaex excels at out-of-core DataFrame operations, lazy evaluation, fast aggregations, efficient visualization of big data, and machine learning on large datasets. Apply when users need to work with large CSV/HDF5/Arrow/Parquet files, perform fast statistics on massive datasets, create visualizations of big data, or build ML pipelines that do not fit in memory.

中文简介正在准备中。当前显示原文说明。

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.

中文简介正在准备中。当前显示原文说明。

Authors reproducible bioinformatics pipelines with Snakemake - rules wired by output-file pattern, wildcards and expand() for sample fan-out, checkpoints for runtime-unknown outputs, resource/retry escalation, and conda/container software deployment on HPC and cloud. Use when deciding rule-based (Snakemake) vs channel/dataflow (Nextflow) authoring; wiring rules by OUTPUT-file pattern rather than imperative order; using wildcards + expand() for sample fan-out and constraining them to stop silent mis-routing; adding checkpoints when the set of outputs is unknown until a step runs (dynamic DAG); diagnosing why a job reran (or did not) under the mtime-plus-provenance trigger set; escalating memory on retry for OOM-killed jobs; and porting a Snakemake 7 `--cluster`/remote-provider command to the Snakemake 8+ executor-plugin and storage-plugin model (snakemake-executor-plugin-slurm) with `--software-deployment-method`.

中文简介正在准备中。当前显示原文说明。

Install, set up, verify, update, pin, uninstall, or troubleshoot the ToolUniverse plugin on OpenAI Codex. ALWAYS consult this skill for any of those — don't answer from memory, because the exact marketplace name (mims-harvard/ToolUniverse), the "codex plugin marketplace add" then "codex plugin add -m tooluniverse" flow, Codex's startup auto-upgrade behavior, the uvx tooluniverse MCP server, and the API-key env vars are easy to get wrong. Use it whenever someone wants to get ToolUniverse (or "the 1000+ scientific tools" / "the harvard tools") working on Codex, says the Codex plugin or its tools/skills won't load, hits a uvx or MCP-server startup error, asks how Codex updates it, wants to pin or remove it, or finds it running an old tool version — even if they never say the word "plugin". Not for the Claude Code plugin (use tooluniverse-claude-code-plugin), for running research with the tools, or for authoring new tools or skills.

中文简介正在准备中。当前显示原文说明。

Runs and configures curated nf-core community Nextflow pipelines (rnaseq, sarek, atacseq, methylseq, ampliseq, taxprofiler, fetchngs) reproducibly, pinning the pipeline revision with -r and selecting a container engine and institutional config via -profile. Use when deciding to adopt a community pipeline versus author one from scratch; picking a pipeline and pinning its -r revision; selecting -profile test/docker/singularity/conda plus an institutional config from nf-core/configs; building and validating a samplesheet CSV against the pipeline schema (nf-schema); choosing --genome/iGenomes versus custom references; configuring resources and max_memory for SLURM/AWS Batch; using -resume and -stub; and reading MultiQC outputs.

中文简介正在准备中。当前显示原文说明。

Check an externally created PR for coding rule violations against memory/INDEX.md atoms and docs/architecture/code-review-rules.md.

中文简介正在准备中。当前显示原文说明。

Scrap and regenerate the in-flight EF migrations on the current branch as one consolidated migration. Use when migrations have accumulated mid-development cruft (added/removed columns, stacked add-column-to-existing-table fixes, hand-edited SQL that needs to come out, schema changes that should have been one migration but ended up as five), or when the branch has merged main and its migrations are now stuck mid-chain so `dotnet ef migrations remove` is unsafe. Triggers on phrases like 'regen the migrations', 'redo these migrations', 'scrap and regen migrations', 'consolidate the in-flight migrations', 'redo the migration stack', or any time an agent hits the mid-chain situation in `memory/architecture/migration-regen-after-rebase.md`.

中文简介正在准备中。当前显示原文说明。

Automated CLI-based parallel agent execution — spawn subagents via Gemini CLI, coordinate through MCP Memory, monitor progress, and run verification

中文简介正在准备中。当前显示原文说明。

Design AI agents with tools, memory, and reasoning. TRIGGERS - Use when user needs help with ai-agent-builder related tasks.

中文简介正在准备中。当前显示原文说明。

Design multi-step AI agent architectures with tools, memory, and orchestration. TRIGGERS - Use when user wants to design AI agents, multi-agent systems, or autonomous AI workflows.

中文简介正在准备中。当前显示原文说明。

Hermes Agent (Nous Research の AI CLI エージェント)。インストール、 CLI コマンド (chat, model, gateway, config, cron, webhook, mcp, skills, profile, hermes pm)、 Configuration (Terminal Backends, Memory, TTS / STT, Auxiliary Models, Compression, Display, Local Models llama.cpp, Egress / Secrets)、 Features (Tools, Skills, Hooks, Delegation, Kanban, Cron, Voice Mode, Personality, Context Files)、 Messaging Gateway (Platform Support, Bot Mode, Connectors)、Security、Architecture、 Hermes Desktop、Plugin Catalog、Profiles、Developer Guide、Nous Portal (Tool Gateway)。

中文简介正在准备中。当前显示原文说明。

NVIDIA CUDA GPU 並列コンピューティングリファレンス。 CUDA C++ / CUDA Python, kernel, nvcc, Unified Memory, CUDA Graphs, Cooperative Groups, Green Contexts, Driver API, マルチ GPU。 PTX ISA 命令セット, state space, MMA 命令。 Blackwell チューニング, Streaming Multiprocessor, NVLink, Locality Domains, Compute Fabric Transport。 CUTLASS / CuTe DSL / CuTe C++ GEMM, Tensor Core, tcgen05, Task Scheduling, Operator API, nvMatmulHeuristics, nsight-compute, compute-sanitizer。

中文简介正在准备中。当前显示原文说明。

Claude Code (code.claude.com) の CLI 本体リファレンス。 CLAUDE.md memory, settings.json (All settings, settings precedence, managed-settings.json), env-vars, model config, keybindings, statusline, terminal config, cli-reference (-p / --continue / --resume), sessions, checkpointing, worktrees, remote control, channels, deep links, troubleshooting。

中文简介正在准备中。当前显示原文说明。

Claude Managed Agents (platform.claude.com) リファレンス。Agents / Sessions / Environments API, SSE event streaming, event deltas, session threads, session budgets, multiagent orchestration (subagents, dynamic workflows, workflow runs, advisor), deployments (cron), dreams, vaults, memory stores, self-hosted sandboxes, webhooks, web search / fetch ドメイン制限, secure deployment, permission policies, Agent Skills, MCP connector, user profiles。

中文简介正在准备中。当前显示原文说明。

Claude API (platform.claude.com) の tool use / Agent Skills / MCP。 tool definition (JSON Schema), tool_use / tool_result, server tools（code execution, web search, web fetch, memory tool, tool search）, client toolset（computer use, browser use）, Agent Skills (SKILL.md), Skills API GA (/v1/skills), MCP connector, mcp_servers, mcp_tool_listing, MCP tunnels, Tunnels API。

中文简介正在准备中。当前显示原文说明。

Apple Silicon GPU コンピュート／数値計算リファレンス。MSL (address space, atomic, SIMD-group, TensorOps, cooperative_tensor, matmul2d), threadgroup sizing, argument buffer, indirect command buffer, unified memory, GPU counters, MTLTensor, MTL4MachineLearningCommandEncoder, MPS / MPSGraph / MPSMatrix / MPSNDArray / MPSNNGraph, MLX (mx.array, lazy evaluation, grad / vmap, mx.compile, mx.fft, export_function, mlx.nn, distributed)。

中文简介正在准备中。当前显示原文说明。

Design what an employer will have a candidate do during an Otameshi Tenshoku trial: turn a business problem into a bounded paid assignment or an unpaid learning visit, with scope, completion criteria, candidate hours, company staff hours, duration, compensation and change conditions, and check the planned cost against the budget. Use when a hiring manager, recruiter or small-business owner wants to decide what work to offer, how much time it takes, what it costs, or how to adjust a plan when materials are late or the scope grows; do not use to write the public listing (trial-listing-draft), to reply to applicants (applicant-reply), to state market rates, legal, tax or service conditions from memory, or to post, contact or contract on the user's behalf.

中文简介正在准备中。当前显示原文说明。

Draft or review an employer's Otameshi Tenshoku listing so a candidate can decide whether to apply: check item by item that the work, deliverable, out-of-scope tasks, required experience, support, hours, schedule, location, compensation, contract form and payment timing are stated, find open-ended, unpaid-company-work, guarantee, personal-attribute and unverifiable wording, and confirm that the hours, weekly availability, period, rate and budget hold together. Use when a hiring manager, recruiter or small-business owner wants a listing written from a plan, wants an existing draft or published listing reviewed before or after posting, or wants to know why applications are not coming; do not use to design the work itself (trial-work-design), to reply to applicants (applicant-reply), to state market rates, legal, tax or service conditions from memory, or to post or update the listing on the user's behalf.

中文简介正在准备中。当前显示原文说明。

Help job seekers use Otameshi Tenshoku to prepare profiles and applications, clarify trial conditions, plan learning visits or paid trial work, and reflect on the experience. Use when a candidate wants to apply for, prepare for, negotiate the scope of, or review an otameshi work experience; do not use to predict whether they will be hired or to judge their aptitude, to state fees, contract terms, or service conditions from memory, or to submit an application, update a profile, or send a message on their behalf. Employer-side requests belong to the employer skills (trial-work-design, trial-listing-draft, applicant-reply and others).

中文简介正在准备中。当前显示原文说明。

Inventory a person's work history as facts before any document is written, listing each period of employment and what they actually did, splitting every experience into situation, their own actions, outcome, and stated role, and recording what each claim rests on (public record, their own records, someone who was there, or memory alone). Use when a user is starting a job search, needs material for a rirekisho, shokumu-keirekisho, or entry sheet, or wants to find which periods of their history they have never written down; do not use to judge their strengths, aptitude, or market value, to name a job they are suited for, or to write finished application copy.

中文简介正在准备中。当前显示原文说明。