“deep”的搜索结果

Adapt PyTorch training scripts for multiple GPUs or machines with Accelerate. This skill guides code changes, configuration, and launch commands, including mixed precision, which combines numerical formats, and DeepSpeed or FSDP integration for distributed training.

中文简介正在准备中。当前显示英语简介。

• Moving single-GPU training to multiple GPUs
• Configuring mixed precision training
• Setting up DeepSpeed or FSDP

Build and train Mixture of Experts models, which activate selected specialist networks for each input. Use DeepSpeed or HuggingFace workflows to configure routing, balance workloads, distribute experts across GPUs, and optimize inference.

中文简介正在准备中。当前显示英语简介。

• Designing MoE models for a limited compute budget
• Implementing Mixtral-style expert routing
• Adjusting expert workloads and training stability

Plan substantial DeepChat changes with a written design proposal and a clear implementation path. Record decisions, track progress, review the completed change, and select focused validation for features, complex bugs, or architecture updates.

中文简介正在准备中。当前显示英语简介。

• Planning a substantial new feature
• Recording the diagnosis and fix for a complex bug
• Documenting architecture changes or migrations

Add an LLM provider to DeepChat by choosing a suitable API connection method, updating authentication and model configuration, and making focused source changes with tests for the affected behavior.

中文简介正在准备中。当前显示英语简介。

• Adding an OpenAI-compatible provider
• Connecting a provider through an existing transport
• Implementing provider-specific authentication

Send prompts to Novita AI models such as DeepSeek and GLM from the command line. Retrieve answers as readable text or structured JSON, stream responses as they arrive, and check available models and API connectivity.

中文简介正在准备中。当前显示英语简介。

• Send questions to DeepSeek or GLM
• Display responses as they arrive
• Retrieve AI responses as JSON

Guide multi-step coding tasks with deepseek-v4-pro or deepseek-v4-flash in thinking mode. Verify files and symbols, write plans with confirmed file paths and line numbers, and use a verifier agent before changes spanning three or more files.

中文简介正在准备中。当前显示英语简介。

• Verify that files and functions exist before editing
• Write plans with confirmed file paths and line numbers
• Check assumptions before changes spanning three or more files

Search open-access academic papers through DeepXiv, then move from brief summaries to section outlines and individual sections. Find trending papers and focus reading on the parts needed to answer a research question.

中文简介正在准备中。当前显示英语简介。

• Find papers on a research topic
• Read only a paper's Introduction section
• Explore papers trending over the past two weeks

Run an AI vulnerability scan on uncommitted changes, changes against main, or an entire codebase. Filter candidate files, investigate potential security issues, and review findings with severity ratings, with options to revalidate and export results.

中文简介正在准备中。当前显示英语简介。

• Checking uncommitted changes for vulnerabilities
• Reviewing security issues in changes against main
• Scanning an existing codebase

Delegate code review, development, and research from OpenCode to DeepSeek Harness. The skill checks service readiness, tracks each task, and returns reports or proposed code changes for the main agent to review.

中文简介正在准备中。当前显示英语简介。

• Get a second opinion on code or architecture
• Delegate code changes to an external agent
• Assign research tasks and track progress

Prepare and publish DeepChat releases through its project-specific workflow. Update versions and bilingual changelogs, run release checks, manage release branches and tags, and resume unfinished releases before publishing after release PR approval.

中文简介正在准备中。当前显示英语简介。

• Prepare a version update and bilingual changelog
• Check which release steps remain
• Resume an unfinished release after metadata changes

Build molecular property prediction workflows with DeepChem for solubility, toxicity, and related chemistry tasks. Prepare molecular representations, choose compatible models, handle missing labels, and evaluate predictions using data splits designed for the intended scientific question.

中文简介正在准备中。当前显示英语简介。

• Predicting solubility from molecular structures
• Comparing graph models on MoleculeNet datasets
• Checking data splits for information leakage

Set up repeatable security tests for LLM apps with Promptfoo and DeepTeam. Connect them to automated build pipelines, detect renewed vulnerabilities to malicious instructions or data leakage, and compare results across model or prompt changes.

中文简介正在准备中。当前显示英语简介。

• Running automated attack tests on every code change
• Blocking merges when security tests fail
• Comparing security across model or prompt versions

Adapt existing PyTorch training scripts to run across multiple GPUs or machines with Accelerate. This skill covers device setup, distributed training configuration, and mixed precision, which uses lower-precision numbers to reduce memory use.

中文简介正在准备中。当前显示英语简介。

• Adapt a single-GPU training script for multiple GPUs
• Configure mixed precision to reduce training memory use
• Set up training with DeepSpeed or FSDP

Retrieve prior decisions and working context from Nowledge Mem in DeepChat, and save useful lessons for future work. It also supports requested conversation exports and connection troubleshooting while keeping operations within the selected server and Space.

中文简介正在准备中。当前显示英语简介。

• Resume work using context from previous sessions
• Recall earlier decisions and their recorded details
• Save reusable lessons for future conversations

Clean up DeepChat's specification-driven development documents after implementation and validation. Review feature, issue, and architecture files, remove completed or obsolete material when evidence supports deletion, and preserve maintained specifications while checking references.

中文简介正在准备中。当前显示英语简介。

• Remove issue documents after confirming a bug is resolved
• Clear old plans from completed feature work
• Update or remove architecture documents after replacement

Save explicitly requested facts or approved summaries and recall them across DeepChat conversations through Memcode. The skill checks storage status and supports answers with citations, while keeping saving and recall under the user's control.

中文简介正在准备中。当前显示英语简介。

• Save facts for future conversations
• Recall context from previous conversations
• Get answers with citations from saved memories

Operate desktop applications through DeepChat's built-in Computer Use tools. Inspect windows, click controls, enter text, use menus, and arrange windows, checking the visible result after each action.

中文简介正在准备中。当前显示英语简介。

• Entering text and using controls in desktop apps
• Opening files or URLs in an application
• Arranging application windows

Configures deep linking in Flutter using Android App Links (assetlinks.json) and iOS Universal Links (apple-app-site-association) integrated with GoRouter v17.x navigation. Use this skill when setting up HTTPS domain association, handling magic links or email verification links, routing marketing campaign URLs (Google Ads, Facebook Ads) into the app, implementing password reset or invite flows via deep links, fixing deep link verification failures, troubleshooting GoRouter deep link integration, or any time the words 'deep link', 'universal link', 'app link', 'assetlinks', 'AASA', 'magic link', or 'deeplinking' appear in the request. Avoids deprecated custom URL schemes (myapp://).

中文简介正在准备中。当前显示原文说明。

Call Gemini and OpenMaaS models such as Llama, DeepSeek, and Qwen through Google Cloud Agent Platform. Prepare authentication and SDK code, run prompts, configure connection endpoints, and troubleshoot common request errors.

中文简介正在准备中。当前显示英语简介。

• Send a test prompt to Gemini
• Call Llama or DeepSeek with the OpenAI SDK
• Write code to call an existing custom endpoint

Work with Feishu/Lark documents and spreadsheets through DeepChat's plugin tools. Search, summarize, create, or update content, and inspect or edit structured data using the operations available in the active plugin configuration.

中文简介正在准备中。当前显示英语简介。

• Searching and summarizing workspace documents
• Creating documents or appending content
• Inspecting and editing spreadsheet data

Set up and run research that compares evidence across multiple sources, such as market studies, competitor analysis, or fact checks. The skill handles environment checks, installation guidance, research settings, progress reporting, final files, and recovery.

中文简介正在准备中。当前显示英语简介。

• Research markets and compare competitors
• Compare evidence on policies or technologies
• Cross-check facts across multiple sources

Trains and applies base-resolution deep learning models on ChIP-seq / ChIP-nexus / CUT&RUN data. Uses BPNet (Avsec 2021 Nat Genet 53:354; soft motif syntax from ChIP-nexus), chromBPNet (Pampari A et al 2024 bioRxiv; bias-factorized base-resolution profiles), EnFormer (Avsec 2021 Nat Methods 18:1196; 196 kb input, ~100 kb effective receptive field), DeepSEA (Zhou 2015; multi-task CNN), and JASPAR 2026 deep-learning collection (1259 BPNet ChIP models). Performs in silico mutagenesis for variant-effect prediction, DeepLIFT/Grad attribution, and TF-MoDISco motif discovery from attribution scores. Use when predicting variant effects on TF binding, discovering soft motif syntax / cooperativity, integrating ChIP-seq with sequence-only predictions, or applying precomputed JASPAR Deep Learning models to new variants.

中文简介正在准备中。当前显示原文说明。

Trains and applies base-resolution deep learning models on ChIP-seq / ChIP-nexus / CUT&RUN data. Uses BPNet (Avsec 2021 Nat Genet 53:354; soft motif syntax from ChIP-nexus), chromBPNet (Pampari A et al 2025 Nat Genet; bias-factorized base-resolution profiles), EnFormer (Avsec 2021 Nat Methods 18:1196; 196 kb input, ~100 kb effective receptive field), DeepSEA (Zhou 2015; multi-task CNN), and JASPAR 2026 deep-learning collection (1259 BPNet ChIP models). Performs in silico mutagenesis for variant-effect prediction, DeepLIFT/Grad attribution, and TF-MoDISco motif discovery from attribution scores. Use when predicting variant effects on TF binding, discovering soft motif syntax / cooperativity, integrating ChIP-seq with sequence-only predictions, or applying precomputed JASPAR Deep Learning models to new variants.

中文简介正在准备中。当前显示原文说明。

Deep, iterative, multi-source research that explores a topic from multiple perspectives, detects contradictions, and produces a comprehensive vault note with fact-level citations and a confidence map. Use this skill whenever the user says /deep-research, "deep research X", "research X thoroughly", "investigate X deeply", "I need to really understand X", or wants research that goes beyond a quick web search. Also use when the user wants to understand a complex, contested, or multi-faceted topic, compare competing approaches with evidence, or build a comprehensive mental model they can trust for decisions. This is heavier than /research — it searches web, academic papers, Reddit, and the vault, then loops through exploration and critique rounds until the evidence base is solid. When in doubt between /research and /deep-research, prefer this one if the topic is complex, contested, or the user wants depth over speed.

中文简介正在准备中。当前显示原文说明。