Results for “accuracy”

Use when evaluating a newly adapted model's accuracy/eval score. Runs evalscope eval (or the test/srt/run_eval.py fallback) against an OpenAI-compatible endpoint on datasets like gsm8k / aime, producing per-dataset reports and a score. Accuracy only — throughput/latency benchmarks belong to model-speed-benchmark. Triggers on accuracy, eval, evalscope, gsm8k, aime, run_eval — not throughput.

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

Plan and document studies that assess whether an analytical procedure suits its intended use. Select the governing framework, evaluate calibration, accuracy, precision, and detection limits, and compare methods using acceptance criteria defined before data collection.

• Planning analytical validation studies and acceptance criteria
• Evaluating calibration, accuracy, and precision
• Estimating and confirming detection and quantitation limits

Launch an SGLang server, run a quick accuracy check, and capture an execution trace for performance inspection. Return the trace location and generated files, ready to open in Perfetto or Chrome tracing.

• Capture an execution trace of SGLang inference
• Check basic accuracy before profiling
• Record input processing and token generation separately

Reduce LLM weight usage through pruning, which removes selected model parameters. Explore Wanda and SparseGPT for compression without retraining, choose hardware-friendly patterns for faster inference, and compare the compressed model's accuracy with the original.

• Compressing an LLM without retraining
• Comparing pruning methods
• Exploring faster inference on supported hardware

Build probability estimates from historical frequencies and supporting or opposing evidence. Structure forecasts with explicit assumptions and outcome criteria, update them as evidence changes, and assess past predictions using accuracy and confidence calibration.

• Estimating probabilities for technology or economic developments
• Weighing evidence for and against a forecast
• Updating forecast probabilities when new evidence appears

Guides LLM compression through pruning, which removes selected model weights. Covers Wanda and SparseGPT methods without retraining, accuracy comparisons, and structured weight patterns that can accelerate inference on compatible hardware.

• Trying model compression without retraining
• Comparing accuracy before and after pruning
• Choosing between Wanda and SparseGPT

Use when running throughput/latency (serving speed) benchmarks against an already-running model server. Uses sgl_jax.bench_serving to sweep batch-size/concurrency and measure throughput and latency, producing summary.csv and per-point bs_<N>/result.jsonl. Speed only — accuracy/eval belongs to model-accuracy-benchmark. Triggers on throughput, latency, bench_serving, batch-size sweep, tok/s, TTFT, TPOT — not accuracy/eval.

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

Use when thumbnail concepts need real-size, accuracy-first scoring without misleading claims. Not for generating thumbnails or declaring winners that fail the accuracy rubric.

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

End-to-end automated model quantization and accuracy tuning workflow. Use when user asks for automated model quantization and accuracy tuning, e.g. "自动量化", "量化调优", "一键量化", "精度调优", etc. 当用户给出多卡（≥2 卡）卡号时，量化阶段通过 EP 并行适配（msmodelslim-ep-parallel-adaptation）保证多卡 EP 并行，调优主流程仍在本 Skill。

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

Transcribe audio files locally with Whisper CLI, a command-line tool that needs no API key. Produce text or translated subtitles, and choose smaller or larger models to balance processing speed and accuracy.

• Transcribing recorded audio into text
• Creating translated subtitles from audio
• Choosing a model for speed or accuracy

Set up deployment checks for healthcare applications covering clinical decision support, patient information exposure, data integrity, workflows, and integrations. Block releases when critical tests fail, and flag lower workflow or integration pass rates for review.

• Checking an electronic health record application before deployment
• Testing changes to drug interactions or dose validation
• Checking patient data exposure after access control changes

Have an AI agent assess its completed work for accuracy, completeness, clarity, actionability, and conciseness. Receive a scorecard with evidence, an average score, and prioritized improvements tied to the original request.

• Assess code after a multi-file implementation
• Check deliverables after a multi-step workflow
• Evaluate a design document or written analysis

Helps you build similarity search with FAISS to find nearby vectors, numerical representations of data. Covers index selection, saving and loading, GPU use, and tuning search speed, accuracy, and memory consumption for large datasets.

• Building similarity search over large vector datasets
• Choosing indexes for speed, accuracy, and memory use
• Running batch searches on GPUs

Compare README, specification, and product documents with recent git history to find outdated claims or missing features. Make small factual corrections, preserve the existing style, and prepare a pull request for review.

• Review documentation after a release
• Update documentation after major feature merges
• Correct outdated roadmap and support claims

Verify RuView pipeline output against a reference hash, package evidence for others to check, and review accuracy statements. The skill distinguishes measured, claimed, and synthetic results and requires real device logs for hardware validation.

• Checking pipeline output against a reference hash
• Preparing a release evidence bundle for independent verification
• Reviewing accuracy statements in reports and READMEs

Tune AgentDB memory use and search speed through quantization, which stores vectors more compactly, search indexes, caching, and batch operations. Includes configuration examples and benchmarks to help balance accuracy, speed, and storage as datasets grow.

• Reduce memory use in an AgentDB application
• Tune search speed while preserving accuracy
• Plan settings for millions of vectors

Retrieve World Monitor forecasts to review scenario probabilities, changes, and supporting signals by region or topic. Check forecast accuracy records with their audit status, and distinguish unavailable data from an absence of forecasted risks.

• Review conflict forecasts for a specific region
• Check changes in scenario probabilities
• Inspect signals behind a forecast

Evaluate improvements to Archify diagrams using actual outputs and agent work records. Compare accuracy, readability, and total delivery time to investigate layout, connection routing, diagnostics, and authoring guidance, then document what the evidence supports.

• Comparing diagram layouts and connection routes
• Investigating repeated diagram repair attempts
• Testing changes to authoring guidance

Review OpenAI Cookbook documentation and notebook prose for technical accuracy, readable wording, consistent terminology, and publication metadata. Make focused edits, preserve notebook code and outputs, and report unresolved issues with validation results.

• Review documentation before merging a pull request
• Polish notebook Markdown cells
• Check publication metadata and author records

Create agents that learn from interactions and adapt their behavior over time. Monitor learning progress, review efficiency and accuracy metrics, trigger adjustments from feedback, and share acquired knowledge with other agents through dedicated tools.

• Create an agent with initial learning settings
• Monitor learning progress, efficiency, and accuracy
• Adjust agent behavior based on feedback

Tune AgentDB memory use and search speed through quantization, which stores vectors more compactly, search indexes, caching, and batch operations. Includes configuration examples and benchmarks to help balance accuracy, speed, and storage as datasets grow.

• Reduce memory use in an AgentDB application
• Tune search speed while preserving accuracy
• Plan settings for millions of vectors

Review existing Appwrite blog posts for tone, structure, SEO, formatting, and product accuracy. Compare claims with local documentation, identify specific problems, and produce a structured report with suggested rewrites and prioritized fixes.

• Review an Appwrite blog draft before publication
• Audit titles, summaries, and search intent
• Verify feature descriptions and API details against local docs

Design how an AI agent finds and loads tools as needed. Compare discovery approaches, organize large tool collections, and assess tool selection accuracy, token costs, and cache reuse without placing every tool definition in the initial context.

• Designing on-demand loading for many MCP tools
• Comparing tool discovery approaches
• Investigating cache misses after loading tools

Evaluate BMad skills using your agent's configured runner. Compare results with the model alone or another version, grade outputs against criteria, and test whether descriptions trigger correctly, preserving transcripts and evidence for each run.

• Compare a skill with the model alone
• Compare an edited skill with a previous version
• Grade outputs against defined criteria