Results for “kg”

Reverse-engineers a federal RFP we received — given the SOW/PWS and evaluation criteria already in the Theseus KG, reconstructs the CO's hidden decision tree (upstream `sow-pws-builder` 6 scope blocks + 3 intake answers), surfaces hot buttons, ghost language, discriminator hooks, missing-section signals, and CPFF-form / Section-5 / QASP / Key-Personnel traps. USE WHEN the user asks "what scope decisions did the CO already make?", "reverse engineer this RFP", "what hot buttons are hiding in this PWS?", "where are the discriminator hooks?", "did they pick CPFF completion or term form?", "anything suspiciously missing?", or any variant of decoding CO intent. Pulls `requirement`, `deliverable`, `proposal_instruction`, `evaluation_factor`, `clause`, `performance_standard` from the active workspace KG and emits a JSON envelope feeding `proposal-generator`. DO NOT USE FOR proposal prose (`proposal-generator`), pricing (`price-to-win`), clause audit (`compliance-auditor`), or sub SOW (`subcontractor-sow-builder`).

Structured data + RAG. NL2SQL hybrid patterns (text-to-SQL then execute vs embed rows), table embedding strategies (row-level, schema-level, hybrid), semantic layer integration (Cube, dbt metrics), LangChain SQLDatabaseChain, LlamaIndex PandasQueryEngine, safe SQL execution (read-only, sandboxed), schema-aware retrieval. Full PostgreSQL + pgvector hybrid code. USE WHEN: user mentions "tabular RAG", "NL2SQL", "text to SQL", "RAG on tables", "database RAG", "SQL RAG", "semantic layer", "structured data RAG" DO NOT USE FOR: unstructured doc RAG - use `rag-architecture`; metadata filtering only - use `self-querying-retriever`; KG retrieval - use `graph-rag`

Activation-aware weight quantization for 4-bit LLM compression with 3x speedup and minimal accuracy loss. Use when deploying large models (7B-70B) on limited GPU memory, when you need faster inference than GPTQ with better accuracy preservation, or for instruction-tuned and multimodal models. MLSys 2024 Best Paper Award winner.

Quantize large language models to 8/4/3/2/1-bit precision without calibration data, using multiple optimized backends and integrations with HuggingFace Transformers, vLLM, and PEFT/LoRA.

Quantize large language models to 8/4/3/2/1-bit precision without calibration data, using multiple optimized backends for deployment with vLLM or HuggingFace Transformers.

Quantize large language models to 4-bit precision using activation-aware weight quantization, reducing memory footprint and speeding up inference with minimal accuracy loss.

Half-Quadratic Quantization for LLMs without calibration data. Use when quantizing models to 4/3/2-bit precision without needing calibration datasets, for fast quantization workflows, or when deploying with vLLM or HuggingFace Transformers.

Quantize large language models to 4-bit using activation-aware weight quantization, achieving ~3x speedup with minimal accuracy loss for deployment on limited GPU memory.

Half-Quadratic Quantization for LLMs without calibration data. Use when quantizing models to 4/3/2-bit precision without needing calibration datasets, for fast quantization workflows, or when deploying with vLLM or HuggingFace Transformers.

Quantize LLMs to 8/4/3/2/1-bit precision without calibration data, using multiple backends and HuggingFace/vLLM integration.

Activation-aware weight quantization for 4-bit LLM compression with 3x speedup and minimal accuracy loss. Use when deploying large models (7B-70B) on limited GPU memory, when you need faster inference than GPTQ with better accuracy preservation, or for instruction-tuned and multimodal models. MLSys 2024 Best Paper Award winner.

Federal Organizational Conflict of Interest (OCI) sweeper backed by FAR Subpart 9.5 (9.501-9.508) and the active Theseus workspace knowledge graph. USE WHEN the user asks about OCI risk on a bid, organizational conflicts, incumbent conflicts, biased ground rules, unequal access to information, impaired objectivity, or any pre-bid OCI due diligence. Pulls `company`, `incumbent`, `subcontractor`, `customer`, `program_office`, and prior-contract relationships from the workspace KG, classifies each potential conflict into one of the three FAR 9.505 classes (biased ground rules, unequal access, impaired objectivity), and emits a structured findings envelope with mitigation recommendations (firewall, NDA, recusal, novation). DO NOT USE FOR FAR clause coverage audit (use `compliance-auditor`), proposal prose (use `proposal-generator`), competitor research (use `competitive-intel`), or pricing (use `price-to-win`).

Quantize large language models to 4-bit with minimal accuracy loss using GPTQ, enabling deployment of 70B+ models on consumer GPUs with 4× memory reduction and 3-4× faster inference.

Universal transformation λ(ο,K).τ with recursive self-improvement. USE WHEN routing reasoning, validating knowledge graphs, preparing CICM/ANZCA examinations, or when self-improvement of reasoning/architecture/context is required. Routes queries through R0-R3 complexity pipelines, validates topology (η≥target) and governance (KROG), emits per style (Φ), and compounds learnings into knowledge K. Triggers on complexity assessment, multi-step reasoning, examination mode, or /λ invocation.

Recommends weight‑losing and weight‑neutral medications as first‑ and second‑line agents for managing overweight/obese patients with type 2 diabetes. Triggers include when a clinician asks, 'Which diabetes medications will not worsen weight in this obese patient?' or 'Should I avoid sulfonylureas in this patient with T2DM and obesity due to weight gain risk?'

Kernel-Mode Driver Framework (KMDF), the Microsoft-recommended way to write Windows kernel-mode drivers. Covers DriverEntry, EvtDeviceAdd, IRPs and IOCTLs, I/O queues, PnP and Power state machines, IRQL discipline, memory pools, WPP tracing, SAL annotations, and Driver Verifier. USE WHEN: user mentions "KMDF", "WDF kernel", "Windows kernel driver", "DriverEntry", "WdfDriverCreate", "EvtDeviceAdd", "IRP", "IOCTL", "DISPATCH_LEVEL", "PASSIVE_LEVEL", "NTSTATUS", "PoolTag", "WdfRequestComplete" DO NOT USE FOR: UMDF v2 (use `wdf-umdf`), classic WDM-only drivers, file-system filters (FltMgr is a separate framework)

Use this skill to understand the FMCSA cargo securement standards. Covers § 393.100-§ 393.136 general rules, tiedown requirements, working load limits.

Store a key-value pair in the Custodian KV store

Compress large language models using knowledge distillation from teacher to student models. Use when deploying smaller models with retained performance, transferring GPT-4 capabilities to open-source models, or reducing inference costs. Covers temperature scaling, soft targets, reverse KLD, logit distillation, and MiniLLM training strategies.

Calculates first- and second-order option Greeks (Delta, Gamma, Theta, Vega, Rho, Charm, Vanna, Volga) for single contracts or multi-leg positions, with scenario heatmaps and position-level aggregation.

Expert en croissance virale (K-factor, network effects, referral loops, embedded virality)

Erstellt aus einem Schriftsatz und Rohbelegen geordnete, gestempelte Anlagen-PDFs für Fahrgastrechte-Verfahren, inklusive optionalem Sammel-PDF und Übersichtstabelle.

Use this skill when the user asks about Household Goods (HHG) movers — 49 CFR 375 + 376 consumer protection rules, FMCSA HHG-specific authority, mover certification, binding vs non-binding estimates, dispute resolution, weight ticket requirements, valuation coverage, and how HHG operations differ from general freight. Cite 49 CFR 375.

Post-training 4-bit quantization for LLMs with minimal accuracy loss. Use for deploying large models (70B, 405B) on consumer GPUs, when you need 4× memory reduction with <2% perplexity degradation, or for faster inference (3-4× speedup) vs FP16. Integrates with transformers and PEFT for QLoRA fine-tuning.

COYO-700M: Image-Text Pair Dataset

Quantizes LLMs to 8-bit or 4-bit for 50-75% memory reduction with minimal accuracy loss. Use when GPU memory is limited, need to fit larger models, or want faster inference. Supports INT8, NF4, FP4 formats, QLoRA training, and 8-bit optimizers. Works with HuggingFace Transformers.

Use this skill when the user asks about refrigerated (reefer) trucking — cold chain compliance, FDA temperature recording requirements, Sanitary Transportation Rule (STR / 21 CFR 1.900-1.934), commodity-specific temperature ranges, refrigeration unit operation, temperature recorder calibration, and how a load is challenged if temperature deviates. Cite 21 CFR Part 1 Subpart Q.

Post-training 4-bit quantization for LLMs with minimal accuracy loss. Use for deploying large models (70B, 405B) on consumer GPUs, when you need 4× memory reduction with <2% perplexity degradation, or for faster inference (3-4× speedup) vs FP16. Integrates with transformers and PEFT for QLoRA fine-tuning.

Post-training 4-bit quantization for LLMs with minimal accuracy loss. Use for deploying large models (70B, 405B) on consumer GPUs, when you need 4× memory reduction with <2% perplexity degradation, or for faster inference (3-4× speedup) vs FP16. Integrates with transformers and PEFT for QLoRA fine-tuning.

Session-end memory garbage collector. Promotes valuable L2 session facts to L1 atomic memory, wipes L2, and rotates oversized audit logs. Run at end of session to prevent context inflation and storage bloat.

Use this skill when the user asks about state-specific trucking taxes BEYOND IFTA — NY Highway Use Tax (HUT), KY Kentucky Intrastate Tax (KIT), Oregon Weight-Mile Tax, New Mexico Weight-Distance Tax, Connecticut Highway Use Fee, or other state-level road-use taxes for commercial vehicles. Covers thresholds, rates, registration, filing cadence, and how these differ from IFTA. Reference state DOT/DOR resources.