Skill UI

TensorBoard delivers a unified interface to track training metrics, debug network behavior, compare experiments, inspect model graphs, and profile performance for PyTorch or TensorFlow workflows.

Assists with tensorboard visualizer tasks for ML training by guiding setup, best practices, and experiment tracking whenever tensorboard visualization is mentioned, producing production-ready configs and validated outputs.

torchforge is Meta's advanced PyTorch-native library designed for agentic Reinforcement Learning (RL). It fundamentally separates core RL algorithms from complex distributed infrastructure concerns. This allows researchers to rapidly prototype and experiment with novel RL methods (like GRPO) using clean, PyTorch-native abstractions, while automatically handling large-scale tasks like distributed training across thousands of GPUs using Monarch and TorchTitan.

Automates guidance on train/test split workflows inside ML training, surfacing best-practice patterns, code snippets, and validation tips for PyTorch, TensorFlow, and sklearn projects.

Provides production-ready Megatron-Core workflows for training 2B–462B parameter LLMs with tensor/pipeline/expert/context parallelism, maximizing GPU efficiency on H100 clusters.

This comprehensive workflow guides users through the end-to-end process of utilizing cloud-based GPU computing resources on Vast.ai. It covers searching for ideal GPU offers based on specs and price, provisioning an instance, running complex training or inference jobs via SSH, collecting model artifacts, and crucially, destroying the instance to prevent unexpected billing. Ideal for deep learning training and research.

This comprehensive script suite provides methods to optimize GPU cloud computing expenses on Vast.ai. It guides users through selecting the most cost-efficient GPU for their workload, leveraging interruptible (spot) instances, and implementing critical lifecycle management safeguards. Features include detecting idle compute resources, auto-destroy timers, and calculating total cost burn rates, ensuring AI model training and inference jobs stay within budget.

A comprehensive guide to managing training data and model artifacts securely on cloud GPU instances (e.g., Vast.ai). It covers best practices for data transfer (SCP, compressed, cloud staging), implementing AES-256 encryption, managing checkpoints to S3, and performing secure cleanup before instance destruction, ensuring compliance and data integrity.

A comprehensive guide for deploying Machine Learning training jobs and inference services on the Vast.ai GPU cloud. This skill covers the entire deployment lifecycle, including optimizing Docker images for performance, creating automated provisioning scripts using the Vast.ai CLI, implementing robust data transfer strategies (SCP, rsync), and performing post-deployment health checks.

This guide details the production-grade reference architecture for running complex GPU compute workflows on Vast.ai. It covers the foundational three-tier pattern—Orchestrator, GPU Workers, and Artifact Storage—providing blueprints for job queue management, fault-tolerant training pipelines, and robust checkpointing. Use this when designing scalable and reliable ML/AI training systems.

Guides training large language models with Volcano Engine's verl library, covering PPO/GRPO workflows, distributed backends, multi-turn rollouts, reward configs, critic-enabled PPO, and production deployments using Ray workers.

Automates wandb experiment logger tasks within ML training, providing guidance, best practices, and code or configuration examples plus validation to keep data prep, training, hyperparameter tuning, and experiment tracking consistent.