skills.rest
dailycafi
dailycafiCOMMUNITY
Explore all skills by this author →
𝕏@dailycafiGitHub

ml-training

Provide PyTorch training recipes for LLMs, vision, and diffusion models.

4|Updated Mar 8, 2026
Search Tags:#debugging#optimization#llm#training#pytorch#neural networks
One-click install
npx skills add https://github.com/dailycafi/ml-training-skill --skill ml-training-dailycafi
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: ml-training
Source: https://github.com/dailycafi/ml-training-skill/tree/main
Command: npx skills add https://github.com/dailycafi/ml-training-skill --skill ml-training-dailycafi

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill provides battle-tested PyTorch training patterns and best practices, helping you train, fine-tune, and debug neural networks more effectively across various domains.

Core Features & Use Cases

  • Comprehensive Training Recipes: Covers LLMs, computer vision, diffusion models, and more.
  • Optimization Strategies: Includes guidance on optimizers (AdamW, Muon), LR scheduling, and mixed precision.
  • Debugging & Performance: Offers checklists and techniques for troubleshooting common issues like loss spikes, OOM errors, and slow training.
  • Use Case: You're training a new Transformer model for text generation and encounter NaN gradients. You can consult this Skill for debugging steps and optimal hyperparameter settings.

Quick Start

Use the ml-training skill to get best practices for fine-tuning a vision transformer model.

Frequently Asked Questions about ml-training

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

FAQPage Schema
How do I debug NaN gradients and loss spikes during PyTorch model training?▼

To debug NaN gradients and loss spikes during PyTorch model training, apply targeted debugging checklists to adjust optimizer configurations, learning rate scheduling, and mixed precision settings for stable neural network optimization.

What is the best way to configure optimizers and learning rate scheduling for fine-tuning LLMs?▼

The best way to configure optimizers and learning rate scheduling for fine-tuning LLMs involves applying specialized recipes like AdamW or Muon optimizers alongside advanced learning rate scheduling to optimize neural network architecture effectively.

How do I resolve out of memory errors when training diffusion models in PyTorch?▼

To resolve out of memory errors when training diffusion models in PyTorch, implement memory optimization techniques and mixed precision training to reduce VRAM usage while maintaining computational efficiency.

Does this provide PyTorch training recipes for specialized domains like biomedical applications?▼

Yes, it provides PyTorch training recipes tailored for specialized domains, addressing specific challenges in biomedical, genomic, and clinical NLP applications to ensure effective neural network architecture selection and training.

How do I start fine-tuning a vision transformer model with PyTorch?▼

To start fine-tuning a vision transformer model with PyTorch, consult comprehensive training recipes that guide architecture selection, optimizer configuration, and mixed precision training to achieve optimal computer vision performance.

Related Skills

Explore More Software Engineering Skills→
hung-phanhung-phan
1

pytorch

Build and train PyTorch deep learning models with custom training loops.

Community
Advanced
ZurybrZurybr

pytorch-model-cli

Extract PyTorch .pth weights and implement native C/C++ inference.

Community
Advanced
tassiovaletassiovale
13

pytorch-lightning

Automate deep learning model training and evaluation with PyTorch Lightning.

Community
Advanced
hhhi21ghhhi21g

pytorch-lightning

Train PyTorch models with a high-level API and distributed training support.

Community
Advanced