ml-training-recipes

Name: ml-training-recipes
Availability: InStock
Author: OpenCoven

Provide PyTorch training recipes for neural network optimization.

20|25|Updated May 30, 2026

Search Tags:#optimization #PyTorch #neural networks #large language models #training recipes

Ready to run

npx skills add https://github.com/OpenCoven/coven-cave --skill ml-training-recipes-opencoven

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: ml-training-recipes
Source: https://github.com/OpenCoven/coven-cave/tree/main/marketplace/craft-sources/alchemists-crucible/ml-training-recipes
Command: npx skills add https://github.com/OpenCoven/coven-cave --skill ml-training-recipes-opencoven

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

This Skill provides battle-tested PyTorch training recipes to solve common challenges in neural network training, from handling large-scale datasets to optimizing model performance.

Core Features & Use Cases

Domain-Specific Recipes: Offers recipes for LLMs, vision, diffusion, medical imaging, and more.
Optimization Techniques: Covers training loops, optimizer selection, learning rate scheduling, mixed precision, and debugging.
Use Case: If you're struggling with training a large language model and encountering loss spikes or out-of-memory errors, this Skill provides strategies and best practices to address these issues.

Quick Start

Use the ml-training-recipes skill to apply the 'Chinchilla rule' for compute-optimal training of large language models.

Frequently Asked Questions about ml-training-recipes

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

FAQPage Schema

How do I fix loss spikes and out-of-memory errors when training large language models in PyTorch?▼

To fix loss spikes and out-of-memory errors during large language model training, apply PyTorch optimization recipes like mixed precision and learning rate scheduling. These battle-tested strategies stabilize training loops and reduce memory consumption.

What is the Chinchilla rule for compute-optimal training of neural networks?▼

The Chinchilla rule is a compute-optimal training strategy dictating the specific ratio of model parameters to training tokens. PyTorch training recipes apply this rule to large language models to maximize performance for a given compute budget.

How do I implement mixed precision and learning rate scheduling in a PyTorch training loop?▼

To implement mixed precision and learning rate scheduling in a PyTorch training loop, apply optimization recipes that configure automatic scaling and dynamic rate adjustment. These techniques accelerate computation and ensure reliable neural network convergence.

Do I need PyTorch and torch-geometric installed to use these training recipes for diffusion models?▼

Yes, you need PyTorch version 2.0.0 or higher installed to use these training recipes for diffusion models. Specific recipes also require the torch-geometric dependency for handling specialized graph-based neural network architectures.

What are the best practices for debugging neural network training performance across different domains?▼

Best practices for debugging neural network training performance include applying battle-tested recipes for optimizer selection, gradient tracking, and learning rate tuning. These PyTorch strategies address domain-specific issues in vision, diffusion, and medical imaging models.