skills.rest
jonnabio
jonnabioCOMMUNITY
Explore all skills by this author →
𝕏@jonnabioGitHub

ml-training-recipes

Provide PyTorch training recipes, optimizer guidance, and debugging strategies.

6|3|Updated Jan 29, 2026
Search Tags:#debugging#optimization#PyTorch#recipes#training#experiment management
One-click install
npx skills add https://github.com/jonnabio/ace-framework --skill ml-training-recipes-jonnabio
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: ml-training-recipes
Source: https://github.com/jonnabio/ace-framework/tree/main/.ace/packs/ai-research/ml-training-recipes
Command: npx skills add https://github.com/jonnabio/ace-framework --skill ml-training-recipes-jonnabio

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires torch>=2.0.0, and includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

This Skill provides a comprehensive set of battle-tested PyTorch training recipes and insights to help you optimize your training process and achieve better results.

Core Features & Use Cases

  • Training Recipes: Access a curated collection of training recipes for various domains like LLMs, vision, diffusion, medical imaging, and genomics.
  • Optimizer Selection: Guidance on selecting the appropriate optimizer for different parameter types and use cases.
  • Learning Rate Scheduling: Tips and configurations for various learning rate scheduling methods.
  • Mixed Precision & Compilation: Best practices for mixed precision training and model compilation.
  • Memory & Performance: Optimization strategies for improving memory and performance.
  • Hyperparameter Search: Guidance on how to systematically search for the best hyperparameters.
  • Debugging Checklist: A comprehensive checklist for debugging training issues like loss spikes, slow training, and silent failures.
  • Experiment Management: Instructions for tracking experiments and evaluating metrics.

Quick Start

Use the ml-training-recipes skill to get a summary of recommended training configurations for a given task.

Frequently Asked Questions about ml-training-recipes

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I optimize PyTorch training speed and memory usage?▼

You can optimize PyTorch training by applying curated recipes for mixed precision, model compilation, and memory management, which directly improve training speed and reduce memory overhead.

What is the best way to select an optimizer and learning rate schedule for neural networks?▼

Selecting the right optimizer and learning rate schedule involves using targeted recipes that match your specific parameter types and use cases, ensuring robust training and faster convergence.

How do I debug PyTorch training issues like loss spikes and slow training?▼

Debugging training issues like loss spikes and silent failures requires following a comprehensive debugging checklist that systematically isolates slow training and instability problems.

Does this PyTorch training recipes collection work with any neural network architecture?▼

Yes, these PyTorch training recipes apply to any project using PyTorch for training neural networks, providing specific configurations for domains like LLMs, vision, and diffusion models.

Do I need PyTorch installed to use these training recipes and configurations?▼

Yes, you need PyTorch version 2.0.0 or higher installed, along with associated libraries, to execute the training recipes and apply the optimization guidelines.

How do I systematically search for the best hyperparameters during PyTorch training?▼

Systematic hyperparameter search uses provided guidelines to track experiments and evaluate metrics, helping you find the best configurations for improving training accuracy.

Related Skills

Explore More Software Engineering Skills→
OpenCovenOpenCoven
20

ml-training-recipes

Provide PyTorch training recipes for neural network optimization.

Official
Advanced
andrewblockernstandrewblockernst

flow-nexus-neural

Train and deploy neural networks in distributed E2B sandboxes.

Community
Advanced
aayushsoamaayushsoam
1

huggingface-accelerate

Automate device placement and mixed precision for PyTorch distributed training.

Community
Advanced
zeronx798zeronx798

weights-and-biases

Track machine learning experiments, hyperparameters, and model artifacts centrally.

Community
Advanced