Skill UI

This skill provides comprehensive monitoring capabilities for GPU instances running on Vast.ai. It collects key metrics including GPU utilization, instance uptime, temperature, and accumulated costs. Use this when setting up operational monitoring dashboards, configuring critical alerts (e.g., idle GPUs, high temperature, budget overruns), or automating resource usage tracking in a cloud computing environment.

Modal delivers cloud-native serverless Python with on-demand GPUs, persistent storage, and web endpoints to deploy ML workloads, batch jobs, and APIs without writing Dockerfiles.

A comprehensive skill for fine-tuning and training large language and vision models using advanced techniques like TRL (SFT, DPO, GRPO) or Unsloth. This process runs entirely on managed Hugging Face cloud GPUs, eliminating the need for local GPU setup. It supports modern training workflows, model conversion (GGUF for local deployment), and ensures seamless artifact storage to the Hugging Face Hub.

This comprehensive tool manages the complete lifecycle of ML experiments, providing robust deployment capabilities across diverse environments. It supports running training jobs on local GPUs, dedicated remote servers, Vast.ai, and serverless platforms like Modal. The workflow handles environment detection, code synchronization (rsync/git), dependency pinning, and integrates advanced logging tools (e.g., W&B), ensuring reliable, scalable, and reproducible ML training runs.

This skill provides a managed cloud environment for training and fine-tuning advanced vision models. Users can build and customize models for object detection (e.g., DETR, YOLOS), image classification, and segmentation (SAM/SAM2). By leveraging Hugging Face Jobs, training is executed on powerful cloud GPUs without requiring local GPU setup, and all resulting models are automatically saved to the Hugging Face Hub.

Modal is a serverless cloud platform designed for executing Python-based AI/ML workloads. It provides on-demand access to powerful GPUs (e.g., H100, A100), scalable serverless functions, and custom containerization. Use cases include deploying ML models, running large-scale training/inference jobs, creating web APIs, and scheduling data pipelines, all defined entirely in code.

A comprehensive guide for optimizing cloud GPU expenditures on CoreWeave. Learn strategies like right-sizing instances, implementing scale-to-zero for development environments, and using quantization techniques (e.g., AWQ, GPTQ) to fit large models onto smaller, cost-effective GPUs, ensuring maximum efficiency and budget adherence for AI/ML workloads.

This skill guides users through building, deploying, and maintaining machine learning applications on Hugging Face Spaces. It provides comprehensive knowledge on selecting the optimal SDK (Gradio, Docker, Static), choosing appropriate hardware resources (ZeroGPU, dedicated GPUs, CPU-basic), and following best practices to ensure ML applications are deployed efficiently and successfully on the platform.