1,022 件 ・ 関連度順
概要と使いどころ
Tokenizadores rápidos otimizados para pesquisa e produção. Implementação em Rust que tokeniza 1GB em menos de 20 segundos. Suporta algoritmos BPE, WordPiece e Unigram. Treine vocabulários customizados, rastreie alinhamentos, gerencie padding/truncagem. Integração perfeita com transformers. Use quando precisar de tokenização de alto desempenho ou treinamento de tokenizadores customizados.
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。
Ajuste fino de LLMs usando reinforcement learning com TRL - SFT para ajuste de instruções, DPO para alinhamento de preferências, PPO/GRPO para otimização de recompensas e treinamento de modelo de recompensas. Use quando precisar de RLHF, alinhar modelo com preferências ou treinar a partir de feedback humano. Funciona com HuggingFace Transformers.
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。
Quantização pós-treinamento em 4-bits para LLMs com perda mínima de precisão. Use para implantar modelos grandes (70B, 405B) em GPUs de consumo, quando você precisa de redução de memória 4× com <2% de degradação de perplexidade, ou para inferência mais rápida (aceleração de 3-4×) vs FP16. Integra com transformers e PEFT para fine-tuning QLoRA.
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。
Quantização Half-Quadratic para LLMs sem dados de calibração. Use ao quantizar modelos com precisão 4/3/2-bit sem necessidade de conjuntos de calibração, para workflows de quantização rápida, ou ao fazer deploy com vLLM ou HuggingFace Transformers.
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。
Quantiza LLMs para 8-bit ou 4-bit com redução de memória de 50-75% e perda mínima de acurácia. Use quando a memória GPU é limitada, precisa ajustar modelos maiores ou quer inferência mais rápida. Suporta formatos INT8, NF4, FP4, treinamento QLoRA e otimizadores 8-bit. Funciona com HuggingFace Transformers.
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。
Ajuste fino com eficiência de parâmetros para LLMs usando LoRA, QLoRA e 25+ métodos. Use ao fazer fine-tuning de modelos grandes (7B-70B) com memória GPU limitada, quando precisa treinar <1% dos parâmetros com perda mínima de precisão, ou para serviços multi-adapter. Biblioteca oficial do HuggingFace integrada ao ecossistema transformers.
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。
Estenda janelas de contexto de modelos transformer usando RoPE, YaRN, ALiBi e técnicas de interpolação de posição. Use ao processar documentos longos (32k-128k+ tokens), estender modelos pré-treinados além dos limites de contexto originais ou implementar codificações posicionais eficientes. Cobre embeddings rotativos, vieses de atenção, métodos de interpolação e estratégias de extrapolação para LLMs.
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。
Transforms data into compelling visual narratives by applying narrative structure, annotation techniques, scrollytelling patterns, and honest framing to data journalism, presentations, and infographics. Use when creating data-driven articles or reports, designing infographics with narrative, building scrollytelling experiences, annotating charts to guide interpretation, or when user mentions data storytelling, presentation design, annotated chart, narrative visualization.
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。
Agent Media OS, the single skill for every media need in a HyperFrames project. Resolve BGM, SFX, image, icon, brand logo, voice, color grade, or LUT into a frozen local file or paste-ready block + ledger record (one verb, `resolve`); generate via TTS / music / image models when the catalog misses; produce voiceover, transcription, captions, and background removal through one shared audio engine; operate on media (cut / reframe / transform); and reuse assets across projects. Also use for vague feedback that real footage looks dark, flat, boring, should feel retro/camcorder/print/ASCII, needs privacy, or needs a media reveal.
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。
Generate or edit raster images when the task benefits from AI-created bitmap visuals such as photos, illustrations, textures, sprites, mockups, or transparent-background cutouts. Use when Codex should create a brand-new image, transform an existing image, or derive visual variants from references, and the output should be a bitmap asset rather than repo-native code or vector. Do not use when the task is better handled by editing existing SVG/vector/code-native assets, extending an established icon or logo system, or building the visual directly in HTML/CSS/canvas.
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。
VRIO resource analysis combined with Dynamic Capabilities assessment. Use this skill to identify real competitive advantages (Barney VRIO framework), then test whether the organisation can sense new opportunities, seize them, and transform itself to stay ahead (Teece dynamic capabilities). Runs in three modes — full diagnostic for deep first-use sessions, review mode for quick re-assessment, and alert triggers for defining early-warning signals. Updates company-context.md after every run. Triggers on questions about competitive advantage, defensibility, resource assessment, why competitors are gaining ground, organisational agility, or whether advantages will hold. Based on practice by Eterdis (eterdis.com).
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。
Plan and execute Convex schema migrations safely, including adding fields, creating tables, and data transformations. Use when schema changes affect existing data.
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。
Plan and execute Convex schema migrations safely, including adding fields, creating tables, and data transformations. Use when schema changes affect existing data.
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。
Transform audio recordings into professional Markdown documentation with intelligent summaries using LLM integration
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。
Transform Claude Code into a fully autonomous agent system with persistent memory, scheduled operations, computer use, and task queuing. Replaces standalone agent frameworks (Hermes, AutoGPT) by leveraging Claude Code's native crons, dispatch, MCP tools, and memory.
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。
Integracao completa com Amazon Alexa para criar skills de voz inteligentes, transformar Alexa em assistente com Claude como cerebro (projeto Auri) e integrar com AWS ecosystem (Lambda, DynamoDB, Polly, Transcribe, Lex, Smart Home).
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。
Detects prompt injection attacks targeting LLM-based applications using a multi-layered defense combining regex pattern matching for known attack signatures, heuristic scoring for structural anomalies, and transformer-based classification with DeBERTa models. The detector analyzes user inputs before they reach the LLM, flagging direct injections (system prompt overrides, role-play escapes, instruction hijacking) and indirect injections (encoded payloads, multi-language obfuscation, delimiter-based escapes). Based on the OWASP LLM Top 10 (LLM01:2025 Prompt Injection) and Simon Willison's prompt injection taxonomy. Activates for requests involving prompt injection detection, LLM input sanitization, AI security scanning, or prompt attack classification.
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 systematic threat hunt hypothesis framework that transforms threat intelligence, attack patterns, and environmental data into testable hunting hypotheses.
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。
Transform client hearing sheets and meeting notes into structured requirements documents. Use when converting raw hearing data (Japanese/English) into formal requirements, generating traceability matrices, identifying gaps/ambiguities, and mapping requirements to WBS items. Triggers include "hearing to requirements", "meeting notes to specs", "requirements traceability", "gap analysis for requirements", or requests involving hearing sheet 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。
Refactor Swift code toward declarative functional programming with immutability, value types, pure functions, transformations, and closure-based APIs. Use when asked for functional refactoring or `/refactor:functional`.
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。
End-to-end data pipeline for empirical research: fetch economic data from APIs (FRED, World Bank, IMF, BLS, OECD, Yahoo Finance), clean and transform raw data, construct strategy-specific variables, and validate panel structure. Use when asked to fetch data, download data, clean data, merge datasets, prepare analysis-ready data.
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。
Signpost (not a knowledge pack) for OMG modelling specifications — UML, SysML, BPMN, UAF, SoaML, MOF, OCL, XMI, DDS. Contains NO source content: it only points to where the official specs live and what each is for. Use when you need an OMG modelling/architecture specification — the OMG licence forbids redistributing or transforming the text, so these cannot be reconstituted into a pack; download them from OMG directly.
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。
Knowledge base from the DoD Digital Engineering Strategy (2018, OUSD/ODASD(SE)). Use for the Department of Defense's five digital engineering goals — (1) formalize the development, integration, and use of models; (2) provide an enduring authoritative source of truth; (3) incorporate technological innovation; (4) establish supporting infrastructure and environments; (5) transform the culture and workforce — plus their focus areas, the document-to-model and design-build-test→model-analyze-build shifts, model formalisms/provenance, governance and access control of the authoritative source of truth, and the coordinate→plan→pilot→sustain rollout. A vision-and-policy strategy, deliberately non-prescriptive. Does NOT provide a how-to method, tool tutorials, MBSE/SysML mechanics, or implementation-plan content; thin on metrics, acquisition-policy detail, and step-by-step procedure. Excludes the CAC-gated Digital Engineering Body of Knowledge (DEBoK).
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。
Transform a supplied photo into a charming cartoon pixel-art scene with recognizable subjects, source-derived colors and an optional short life-save caption. Use for 卡通像素风, 像素生活存档, retro game scene portraits, pets or travel-photo pixel illustrations.
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。