Results for “ray”

Plan and implement native RayTK (t3kt/raytk) raymarching/SDF toolkit integration into tdmcp — beyond the existing package-manager staging, into AI-driven RayTK operator (ROP) node graphs. Use whenever the user wants to build, plan, design, implement, continue, fix, QA, or ship RayTK support: create/wire RayTK ROPs, a RayTK operator knowledge catalog, a RayTK-native raymarch/SDF scene tool, RayTK recipes, or the RayTK setup/doctor flow — including follow-ups (re-run, continue, refresh after a new RayTK release, re-QA, docs). Also triggers on 'raytk', 'raymarching toolkit', 'SDF toolkit', 't3kt', 'montar um build para o raytk', 'integrar o raytk'. This owns RayTK integration specifically; use it before the generic tdmcp-pipeline for this feature. Simple questions about RayTK can be answered directly.

The English summary is being prepared. Showing the original description.

Helps adapt machine learning training to multiple GPUs and machines with Ray Train. Covers resource allocation, recovery from saved training states, and Ray Tune searches that compare settings such as learning rate and batch size.

• Adapt existing PyTorch training to multiple GPUs
• Run HuggingFace training across multiple machines
• Compare learning rates and batch sizes with Ray Tune

Adapt machine learning training to run across multiple GPUs or machines with Ray Train. Covers changes to existing code, comparing training settings with Ray Tune, and saving checkpoints so training can resume after failures.

• Adapting existing PyTorch code for multiple GPUs
• Training HuggingFace models across multiple machines
• Comparing learning rates and batch sizes with Ray Tune

Search BioStudies and ArrayExpress by keywords or retrieve study information using an accession, a study’s registration identifier. Returns concise summaries and can save complete API responses to a file for further inspection.

• Search for studies related to RNA
• Find single-cell studies in ArrayExpress
• Retrieve study details by accession

Work with large scientific arrays by storing data in smaller blocks and reading selected portions. Guides compression, cloud storage, format migration, and NumPy, Dask, or Xarray integration while preserving dimensions, units, and other metadata.

• Reading selected regions of large scientific arrays
• Choosing data blocks and compression settings
• Working with scientific arrays stored on S3 or GCS

Review cross-origin isolation, a browser security mode required by features such as SharedArrayBuffer. Check response headers and external resources, identify integrations that could break, and verify that isolation actually works in the browser.

• Checking isolation requirements for SharedArrayBuffer
• Auditing external resources before enabling COEP
• Reviewing effects on popups and third-party embeds

Build data processing pipelines for machine learning, from preprocessing to batch predictions. Uses Ray Data to handle datasets larger than memory, distribute work across CPUs and GPUs, and feed data into training frameworks.

• Preprocessing datasets larger than memory
• Running batch predictions across multiple GPUs
• Preparing image and tabular data for training

Build data processing pipelines for machine learning, from preprocessing to batch predictions. Uses Ray Data to handle datasets larger than memory, distribute work across CPUs and GPUs, and feed data into training frameworks.

• Preprocessing datasets larger than memory
• Running batch predictions across multiple GPUs
• Preparing image and tabular data for training

Fix Native SDK build failures caused by older Zig APIs. Organized by compiler error, this guide shows Zig 0.16 patterns for file access, output, lists, processes, environment variables, clocks, networking, and build configuration.

• Migrating older Zig code to Zig 0.16
• Fixing file access and output errors
• Updating ArrayList allocation calls

Analyze memory dumps with Volatility3 to find heap spraying, an attack that fills memory with attacker-controlled data. Examine suspicious allocations and repeated byte patterns, extract suspicious code, and report affected processes and memory regions in JSON.

• Investigating suspected heap spraying in a memory dump
• Extracting suspicious code from memory
• Organizing evidence for detection rule development

推特中文圈 AI 自媒体博主 (推特(X) 中文圈 AI 自媒体博主 (从业者视角) — 在 X/推特 用中文做 AI 内容的自媒体博主，重点是**怎么写推文/thread、怎么制作内容**。覆盖: (a) 选题 — AI 模型发布解读 / prompt 技巧 / AI 工具实测 / AI 出海与独立开发(build in public) / 英文一手编译翻译 / 信息差 / AI 资讯快讯 / 深度长推; (b) 文案与内容结构(核心) — 单条推(开头钩子/信息密度/短句断行/配图截图/数据图)、长推与 thread(首条钩子/编号/逻辑链/结尾关注 CTA)、引用锐评(quote tweet)、排版可读性、配图(ray.so/carbon 代码图/产品截图/对比图); (c) 内容制作工作流 — 找信息源(英文一手/arXiv/官方/Reddit/HN)→选题→编译总结加洞察→写推/thread→配图→发布→互动维护; (d) AI 工具用于内容生产 — 英文一手编译总结(GPT/Claude/Kimi)、配图(ray.so/carbon/截图美化)、thread 工具(Typefully/Hypefury)、翻译、排程分析; (e) 平台机制与涨粉 — X 算法时间线(转评赞/互动权重/外链降权)、蹭热点借势、互动与回复、build in public 涨粉、英文区搬运到中文区的信息差红利; (f) 变现(次要) — 广告 / 知识星球 / 社群 / 付费课程 / 咨询 / 导流。学派分歧: 编译搬运(英文一手翻译) vs 原创洞察、快讯资讯 vs 深度长推、锐评玩梗 vs 严肃干货、build in public vs 纯内容、全 AI 自动发 vs 人工精选。代表生态(宝玉/歸藏/向阳乔木/AI 进化论/小互/Gorden Sun 等)。不含: 英文 AI Twitter 圈(聚焦中文圈)、通用 X 增长营销、其他平台、AI 工具技术原理。) Master OS — automated mastery of 推特(X) 中文圈 AI 自媒体博主 (从业者视角) — 在 X/推特 用中文做 AI 内容的自媒体博主，重点是**怎么写推文/thread、怎么制作内容**。覆盖: (a) 选题 — AI 模型发布解读 / prompt 技巧 / AI 工具实测 / AI 出海与独立开发(build in public) / 英文一手编译翻译 / 信息差 / AI 资讯快讯 / 深度长推; (b) 文案与内容结构(核心) — 单条推(开头钩子/信息密度/短句断行/配图截图/数据图)、长推与 thread(首条钩子/编号/逻辑链/结尾关注 CTA)、引用锐评(quote tweet)、排版可读性、配图(ray.so/carbon 代码图/产品截图/对比图); (c) 内容制作工作流 — 找信息源(英文一手/arXiv/官方/Reddit/HN)→选题→编译总结加洞察→写推/thread→配图→发布→互动维护; (d) AI 工具用于内容生产 — 英文一手编译总结(GPT/Claude/Kimi)、配图(ray.so/carbon/截图美化)、thread 工具(Typefully/Hypefury)、翻译、排程分析; (e) 平台机制与涨粉 — X 算法时间线(转评赞/互动权重/外链降权)、蹭热点借势、互动与回复、build in public 涨粉、英文区搬运到中文区的信息差红利; (f) 变现(次要) — 广告 / 知识星球 / 社群 / 付费课程 / 咨询 / 导流。学派分歧: 编译搬运(英文一手翻译) vs 原创洞察、快讯资讯 vs 深度长推、锐评玩梗 vs 严肃干货、build in public vs 纯内容、全 AI 自动发 vs 人工精选。代表生态(宝玉/歸藏/向阳乔木/AI 进化论/小互/Gorden Sun 等)。不含: 英文 AI Twitter 圈(聚焦中文圈)、通用 X 增长营销、其他平台、AI 工具技术原理。: top builders' mental models, tool stack, current workflows, jargon, and where to keep up. Trigger this skill when the user works on 推特(X) 中文圈 AI 自媒体博主 (从业者视角) — 在 X/推特 用中文做 AI 内容的自媒体博主，重点是**怎么写推文/thread、怎么制作内容**。覆盖: (a) 选题 — AI 模型发布解读 / prompt 技巧 / AI 工具实测 / AI 出海与独立开发(build in public) / 英文一手编译翻译 / 信息差 / AI 资讯快讯 / 深度长推; (b) 文案与内容结构(核心) — 单条推(开头钩子/信息密度/短句断行/配图截图/数据图)、长推与 thread(首条钩子/编号/逻辑链/结尾关注 CTA)、引用锐评(quote tweet)、排版可读性、配图(ray.so/carbon 代码图/产品截图/对比图); (c) 内容制作工作流 — 找信息源(英文一手/arXiv/官方/Reddit/HN)→选题→编译总结加洞察→写推/thread→配图→发布→互动维护; (d) AI 工具用于内容生产 — 英文一手编译总结(GPT/Claude/Kimi)、配图(ray.so/carbon/截图美化)、thread 工具(Typefully/Hypefury)、翻译、排程分析; (e) 平台机制与涨粉 — X 算法时间线(转评赞/互动权重/外链降权)、蹭热点借势、互动与回复、build in public 涨粉、英文区搬运到中文区的信息差红利; (f) 变现(次要) — 广告 / 知识星球 / 社群 / 付费课程 / 咨询 / 导流。学派分歧: 编译搬运(英文一手翻译) vs 原创洞察、快讯资讯 vs 深度长推、锐评玩梗 vs 严肃干货、build in public vs 纯内容、全 AI 自动发 vs 人工精选。代表生态(宝玉/歸藏/向阳乔木/AI 进化论/小互/Gorden Sun 等)。不含: 英文 AI Twitter 圈(聚焦中文圈)、通用 X 增长营销、其他平台、AI 工具技术原理。 problems and wants industry-grade thinking, tool selection, or workflow guidance. 触发词：「推特 ai 博主」「推特中文 ai」「x ai 自媒体」「推特 thread」「推特爆款推文」

The English summary is being prepared. Showing the original description.

Create an Estuary derivation that flattens a nested array into one output document per array element. Use for normalising nested JSON, exploding line items, unpacking tags, or producing one-row-per-item tables from arrays. Derivations add complexity and cost — confirm the user wants a derivation before reaching for this (some destinations can `UNNEST` server-side without one). Use when user says "flatten-array derivation", "derivation to explode an array", "transformation to unnest", "derive one row per item", or "flatten nested data with a derivation".

The English summary is being prepared. Showing the original description.

Build a single HTML visualizer for particles, fractals, simulations, or data. It moves heavy calculations into a Web Worker, a background browser process, and draws on a canvas while targeting smooth 60fps rendering.

• Prototyping a particle or physics simulation
• Visualizing Mandelbrot or Julia fractals
• Building an audio or data visualizer

Use Dask to process tables, numerical arrays, and file collections in pieces across CPU cores or machines. Helps choose data structures, size memory-friendly chunks, coordinate dependent tasks, and diagnose slow or memory-heavy workflows.

• Aggregate CSV or Parquet files that exceed available memory
• Process large scientific arrays in chunks
• Clean JSON records and logs before analysis

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.

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

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.

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

Audits arithmetic code for mixed units, decimal precision mistakes, and scaling errors. Adds comments identifying each value’s units and produces a summary of findings, unit definitions, and files reviewed.

• Reviewing token amounts and price conversions in DeFi code
• Checking WAD/RAY scaling errors
• Documenting units and precision in existing formulas

Expert blueprint for physics queries using RayCast, ShapeCast, and DirectSpaceState. Covers hit detection, volume overlap, mouse picking, and high-performance server-side intersection queries. Use when implementing projectiles, LOS, terrain grounding, or AI sensors. Keywords raycast, shapecast, direct_space_state, intersect_ray, intersect_shape, PhysicsRayQueryParameters, collision mask, mouse picking.

The English summary is being prepared. Showing the original description.

Expert blueprint for physics queries using RayCast, ShapeCast, and DirectSpaceState. Covers hit detection, volume overlap, mouse picking, and high-performance server-side intersection queries. Use when implementing projectiles, LOS, terrain grounding, or AI sensors. Keywords raycast, shapecast, direct_space_state, intersect_ray, intersect_shape, PhysicsRayQueryParameters, collision mask, mouse picking.

The English summary is being prepared. Showing the original description.

Fix ClickHouse query errors in Rust when using clickhouse-rs crate. Use when: (1) "string is not valid utf8" errors - typically FixedString columns need CAST to String, (2) "tag for enum is not valid" errors - typically Option<T> fields receiving non-NULL values, (3) Sum/count aggregations returning Float64 but Rust expects u64, (4) LEFT JOIN results with empty strings where Rust expects Option::None, (5) "not enough data, probably a row type mismatches a database schema" - query SELECT returns fewer columns than the Rust Row struct expects (common when multiple query functions share the same struct but one function is missing columns added later), (6) Garbled/corrupted data when INSERTing to FixedString(N) columns using String Rust type - silently corrupts data (no error!) because String adds a length prefix but FixedString expects exactly N raw bytes. Fix: use [u8; N] with #[serde(with = "BigArray")] from serde-big-array, (7) "the trait bound `[u8; 64]: serde::Serialize` is not satisfied" when using [u8; N] arrays in Row structs - need #[serde(with = "BigArray")] annotation from serde-big-array crate.

The English summary is being prepared. Showing the original description.

Agent fallback for CCSPlayerPawn_SprayPaint (auto-generated, category: func). Locate CCSPlayerPawn::SprayPaint in the CS2 server module via IDA Pro MCP and emit a fresh, minimal-unique artifact. The deterministic preprocessor could not resolve this symbol on the current gamever - your job is the re-sign. Trigger: CCSPlayerPawn_SprayPaint, CCSPlayerPawn::SprayPaint

The English summary is being prepared. Showing the original description.

Review JavaScript loops and replace suitable ones with standard array and object methods. Make transformations, filtering, and aggregation easier to understand, with practical explanations and browser checks to confirm that changes preserve the intended behavior.

• Refactoring manual loops into array methods
• Clarifying data transformation, filtering, and aggregation
• Reviewing object transformations and deep copying

Add volumetric underwater god rays to a Three.js scene, marching one drifting swell field along every view ray at half resolution so the light shafts converge on the sun, parallax with the camera and carry rippling caustic streaks, with a shimmering surface underside, a sun that wobbles through the ripples, and the same light dappling fish as they swim through the beams. Use for underwater scenes, oceans seen from below, aquariums, reefs, kelp forests, diving and shipwreck scenes, schools of fish, sunbeams or light shafts through water, caustics, "rays of light shining through the ocean", or watery ripples on a background. Includes a reusable module and the Blue Hour sardine-school demo.

The English summary is being prepared. Showing the original description.

Scalable data processing for ML workloads. Streaming execution across CPU/GPU, supports Parquet/CSV/JSON/images. Integrates with Ray Train, PyTorch, TensorFlow. Scales from single machine to 100s of nodes. Use for batch inference, data preprocessing, multi-modal data loading, or distributed ETL pipelines.

The English summary is being prepared. Showing the original description.