Найдите подходящую возможность для вашего агента.
Production-ready reinforcement learning algorithms (PPO, SAC, DQN, TD3, DDPG, A2C) with scikit-learn-like API. Use for standard RL experiments, quick prototyping, and well-documented algorithm implementations. Best for single-agent RL with Gymnasium environments. For high-performance parallel training, multi-agent systems, or custom vectorized environments, use pufferlib instead.
Cost optimization for AI workloads - model selection, GPU sizing, commitment strategies, and multi-cloud cost management
Deep learning foundations including neural network basics, backpropagation, optimization, regularization, and training best practices.
High-performance reinforcement learning framework optimized for speed and scale. Use when you need fast parallel training, vectorized environments, multi-agent systems, or integration with game environments (Atari, Procgen, NetHack). Achieves 2-10x speedups over standard implementations. For quick prototyping or standard algorithm implementations with extensive documentation, use stable-baselines3 instead.
This skill should be used when working with genomic interval data (BED files) for machine learning tasks. Use for training region embeddings (Region2Vec, BEDspace), single-cell ATAC-seq analysis (scEmbed), building consensus peaks (universes), or any ML-based analysis of genomic regions. Applies to BED file collections, scATAC-seq data, chromatin accessibility datasets, and region-based genomic feature learning.
Efficient AI techniques including model compression, quantization, pruning, knowledge distillation, and hardware-aware optimization for production systems.
Analyze GRPO training runs for learning dynamics and pipeline performance. Use when diagnosing training issues, reviewing Elo progression, checking throughput, or updating experiment results.
This skill should be used when users want to train or fine-tune language models using TRL (Transformer Reinforcement Learning) on Hugging Face Jobs infrastructure. Covers SFT, DPO, GRPO and reward modeling training methods, plus GGUF conversion for local deployment. Includes guidance on the TRL Jobs package, UV scripts with PEP 723 format, dataset preparation and validation, hardware selection, cost estimation, Trackio monitoring, Hub authentication, and model persistence. Should be invoked for tasks involving cloud GPU training, GGUF conversion, or when users mention training on Hugging Face Jobs without local GPU setup.
Defensive techniques using adversarial examples to improve model robustness and security
This skill should be used when optimizing strategy parameters through grid search, random search, or Bayesian optimization. It provides systematic approaches to find optimal parameter combinations while avoiding overfitting through cross-validation and walk-forward methods.
Guide for adversarial machine learning: adversarial examples, data poisoning, model backdoors, and evasion attacks.
Advanced techniques for optimizing LLM fine-tuning. Covers learning rates, LoRA configuration, batch sizes, gradient strategies, hyperparameter tuning, and monitoring. Use when fine-tuning models for best performance.
Data engineering, machine learning, AI, and MLOps. From data pipelines to production ML systems and LLM applications.
Research pipeline for topology-aware GNN representation learning on power grids using the PowerGraph benchmark. Use when (1) building physics-guided GNNs for power flow (PF), optimal power flow (OPF), or cascading failure prediction, (2) implementing self-supervised pretraining for power systems, (3) evaluating cascade explanation fidelity against ground-truth masks, or (4) conducting reproducible ML-for-power-systems research. Triggers include "PowerGraph", "power flow GNN", "OPF surrogate", "cascade prediction", "physics-guided GNN", "grid analytics ML", "power system representation learning".
Use when building AI-native products where user data can fine-tune performance, when static software fails to improve with usage, or when designing products that learn from interaction
Master training pipelines - orchestration, distributed training, hyperparameter tuning
Privacy attacks to extract training data and sensitive information from AI models
