2,262 件 ・ 関連度順
概要と使いどころ
GPU-accelerated data curation for LLM training. Supports text/image/video/audio. Features fuzzy deduplication (16× faster), quality filtering (30+ heuristics), semantic deduplication, PII redaction, NSFW detection. Scales across GPUs with RAPIDS. Use for preparing high-quality training datasets, cleaning web data, or deduplicating large corpora.
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。
High-performance RLHF framework with Ray+vLLM acceleration. Use for PPO, GRPO, RLOO, DPO training of large models (7B-70B+). Built on Ray, vLLM, ZeRO-3. 2× faster than DeepSpeedChat with distributed architecture and GPU resource sharing.
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。
Selects biomarker features from high-dimensional omics data using Boruta all-relevant selection, mRMR, LASSO/elastic-net, and stability selection, while controlling the leakage, irreproducibility, and correlated-feature traps that make most published signatures fail to replicate. Use when identifying candidate biomarkers, deciding between an all-relevant and a minimal-optimal selector, or judging whether a selected gene set is reproducible. For unbiased performance estimation of the resulting model see machine-learning/model-validation; for interpreting a trained model see machine-learning/prediction-explanation.
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。
Builds diagnostic and prognostic classifiers on omics feature matrices with regularized logistic regression, random forest, and gradient-boosted trees, handling the p>>n regime, batch shortcut learning, class imbalance, and probability calibration. Use when building a classifier from expression, methylation, or variant data, choosing an algorithm for high-dimensional small-n data, or diagnosing a suspiciously perfect AUC. For unbiased evaluation see machine-learning/model-validation; for feature selection see machine-learning/biomarker-discovery; for time-to-event outcomes see machine-learning/survival-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。
Optimizes LLM inference with NVIDIA TensorRT for maximum throughput and lowest latency. Use for production deployment on NVIDIA GPUs (A100/H100), when you need 10-100x faster inference than PyTorch, or for serving models with quantization (FP8/INT4), in-flight batching, and multi-GPU scaling.
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。
High-performance vector similarity search engine for RAG and semantic search. Use when building production RAG systems requiring fast nearest neighbor search, hybrid search with filtering, or scalable vector storage with Rust-powered performance.
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。
Fast tokenizers optimized for research and production. Rust-based implementation tokenizes 1GB in <20 seconds. Supports BPE, WordPiece, and Unigram algorithms. Train custom vocabularies, track alignments, handle padding/truncation. Integrates seamlessly with transformers. Use when you need high-performance tokenization or custom tokenizer training.
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。
UMAP dimensionality reduction. Fast nonlinear manifold learning for 2D/3D visualization, clustering preprocessing (HDBSCAN), supervised/parametric UMAP, for high-dimensional 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。
High-performance toolkit for genomic interval analysis in Rust with Python bindings. Use when working with genomic regions, BED files, coverage tracks, overlap detection, tokenization for ML models, or fragment analysis in computational genomics and machine learning applications.
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。
Create high-quality ToolUniverse skills following test-driven, implementation-agnostic methodology.
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。
Drug-combination synergy analysis — quantify whether two drugs together are synergistic, additive, or antagonistic using the standard reference models (Bliss independence, HSA / highest single agent, Loewe additivity, ZIP, and the Chou-Talalay Combination Index). Use when you have measured single-drug and combination effects (inhibition/viability) and need a synergy score. Explains which model to use, what data each one needs, and how to read the score. NOT for looking up pre-computed synergy in a database (use the SYNERGxDB tool / cell-line-profiling skill).
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。
Publication-quality molecular visualization. 2D structure drawings (PNG/SVG), molecule grids with property annotations, scaffold highlighting, protein-ligand interaction diagrams, and interactive 3D views.
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。
Designs and validates quantitative targeted metabolomics assays (MRM/SRM on triple-quadrupole, PRM on high-resolution instruments) to report absolute concentrations. Covers the internal-standard strategy (external cal -> global IS -> standard addition -> stable-isotope-labeled IS), weighted calibration judged by back-calculated %RE not R-squared, ion-ratio quantifier/qualifier confirmation, matrix-effect/recovery characterization, and ICH M10 method validation. Use when quantifying a closed panel of known metabolites with units, building or validating an LC-MS/MS assay, choosing an IS or calibration weighting, or judging whether a reported concentration is trustworthy. For untargeted feature detection see metabolomics/xcms-preprocessing; for group statistics see metabolomics/statistical-analysis; for flux/MID/tracing see metabolomics/isotope-tracing.
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。
Ultrawork - high-quality 5-phase development workflow with 11 review steps out of 17
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。
Write high-converting ad copy for Facebook, Google, LinkedIn, and TikTok ads. TRIGGERS - Use when user wants ad copy, paid media creative, or advertising text for any platform.
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。
Upscale a low-resolution image using AI super-resolution for print or high-DPI display.
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。
Draws debug overlays in a running CARLA world via world.debug — points, lines, arrows, bounding boxes, and text strings, each with a colour and lifetime. Use when the user asks to "draw/mark/highlight a point or box in the world", "visualise a path/waypoints", "label a location", or "show something on the map for debugging". Shapes are an overlay, not actors.
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。
Expert database administrator specializing in high-availability systems, performance optimization, and disaster recovery. Masters PostgreSQL, MySQL, MongoDB, and Redis with focus on reliability, scalability, and operational excellence.
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。
Multi-agent debate orchestration for Architecture Decision Records. Automatically triggers on ADR create/edit/delete. Coordinates architect, critic, independent-thinker, security, analyst, and high-level-advisor agents in structured debate rounds until consensus.
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。
Creates high-quality content (blog posts, tweets, newsletters, documentation) that matches the user's writing style and voice. Performs web research to find citations and supporting evidence. Use when user requests blog posts, marketing content, newsletters, tweets, or any written content that should sound authentic and be well-researched.
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。
Prepare a request to change the terms of a Japanese job offer before accepting it, sorting what the user wants changed (pay components, start date, work style, location, title, probation) by priority, tying each ask to a basis the user can show (the written offer, the job posting, what was said in interviews, their own records), working the timing back from the acceptance deadline, and drafting the message to the employer or recruiting agent. Use when a user has an offer and wants to ask for higher pay, a later start date, more remote days, or other changed terms, or wants to know how to raise it without losing the offer; do not use to predict whether the request will succeed, to set a target figure from market rates, to fabricate a competing offer, to decide whether to accept, or to send anything to the employer.
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。
Triage a GitHub issue that already has the cc-issue-created label and is assumed ready to start. Inspect the comment history first to decide whether human confirmation is needed (cc-need-human-check, highest priority — reserved for issue-level deadlock, or for confirmation items answer-issue-questions already investigated and explicitly marked unanswerable); otherwise close the issue as not needed, delegate unanswered confirmation items in the last comment or the description to answer-issue-questions (cc-answer-issue-questions), refresh a description stale relative to settled comment-history content (cc-update-issue), or move to execution (cc-exec-issue). Any UI-changing issue is routed to the design-first flow (cc-create-ui-design) before execution; the exceptions are an issue whose description already names the .pen design to implement against, and a minor UI change that introduces no new visual decision (no new screen/section, no layout change, no new reusable component, no visual expression outside existing design tokens/components). Design files are never edited in the implementation PR. Dependency checks are out of scope.
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。
Designs and ranks guide RNAs (sgRNAs) for CRISPR-Cas9/Cas12a gene knockout by scanning a target for PAM sites (NGG SpCas9, NNGRRT SaCas9, TTTV Cas12a, NG SpCas9-NG, near-PAMless SpRY), enumerating candidate spacers, applying hard filters (Pol-III TTTT terminator, 5' G, GC), ranking on-target activity with the context-appropriate model (Rule Set 2/Azimuth for U6/lentiviral, CRISPRscan for T7/embryo, DeepHF for high-fidelity variants, DeepCpf1 for Cas12a), and predicting the indel/frameshift outcome (Bae out-of-frame score, inDelphi, FORECasT, Lindel). Use when selecting sgRNAs to knock out a gene, choosing a nuclease/PAM for a constrained locus, picking which exon to target, or shortlisting guides before an off-target check. Off-target specificity, base/prime editing, and HDR donors are separate skills.
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。
Renders Hi-C contact matrices honestly and reproducibly with matplotlib, cooltools, HiCExplorer, pyGenomeTracks, FAN-C, CoolBox, and plotgardener. Covers the raw/ICE-balanced/observed-over-expected transform choice, LogNorm vs symmetric-diverging colormaps with vmax/percentile clipping, resolution-to-feature matching (compartments 100-500kb, TADs 10-40kb, loops 5-10kb), square vs rotated-triangle track-stacking, NaN/white-stripe handling, virtual 4C, APA/saddle/on-diagonal pileups, two-condition side-by-side and log2-ratio maps, and interactive (HiGlass) vs scripted-static publication figures. Use when plotting a contact matrix, choosing a normalization or color scale, building a multi-track Hi-C figure, making a virtual 4C profile, piling up loops/boundaries, or comparing two conditions.
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。