---
title: Level up
description: >-
  Learn enough Lightning to match the level of expertise required by your
  research or job.
canonical: 'https://lightning.ai/docs/pytorch/stable/reference/expertise_levels'
---
# Level up {#level-up}

Learn enough Lightning to match the level of expertise required by your research or job.

## Basic skills {#basic-skills}

Learn the basics of model development with Lightning. Researchers and machine learning engineers should start here.

- [Level 1: Train a model](/docs/pytorch/2.6.6/reference/model/train_model_basic.md) — Learn the basics of training a model.

- [Level 2: Add a validation and test set](/docs/pytorch/2.6.6/level-up/core_skills/basic_level_2.md) — Add validation and test sets to avoid over/underfitting.

- [Level 3: Use pretrained models](/docs/pytorch/2.6.6/reference/advanced/transfer_learning.md) — Learn how to use pretrained models with Lightning

- [Level 4: Enable script parameters](/docs/pytorch/2.6.6/reference/common/hyperparameters.md) — Add parameters to your script so you can run from the commandline.

- [Level 5: Understand and visualize your model](/docs/pytorch/2.6.6/level-up/core_skills/basic_level_5.md) — Remove bottlenecks and visualize your model

- [Level 6: Predict with your model](/docs/pytorch/2.6.6/level-up/core_skills/core_level_6.md) — Use your model for predictions.

## Intermediate skills {#intermediate-skills}

Learn to scale up your models and enable collaborative model development at academic or industry research labs.

- [Level 7: Hardware acceleration](/docs/pytorch/2.6.6/level-up/intermediate/intermediate_level_7.md) — Learn how to access GPUs and TPUs on the cloud.

- [Level 8: Modularize your projects](/docs/pytorch/2.6.6/level-up/intermediate/intermediate_level_9.md) — Create DataModules to enable dataset reusability.

- [Level 9: Understand your model](/docs/pytorch/2.6.6/level-up/intermediate/intermediate_level_10.md) — Use advanced visuals to find the best performing model.

- [Level 10: Explore SOTA scaling techniques](/docs/pytorch/2.6.6/level-up/intermediate/intermediate_level_11.md) — Explore SOTA techniques to help convergence, stability and scalability.

- [Level 11: Deploy your models](/docs/pytorch/2.6.6/level-up/intermediate/intermediate_level_12.md) — Learn how to deploy your models with optimizations like ONNX and torchscript.

- [Level 12: Optimize training speed](/docs/pytorch/2.6.6/level-up/intermediate/intermediate_level_13.md) — Use advanced profilers to mixed precision to train bigger models, faster.

- [Level 13: Run on on-prem clusters](/docs/pytorch/2.6.6/level-up/intermediate/intermediate_level_14.md) — Run on a custom on-prem cluster or SLURM cluster.

## Advanced skills {#advanced-skills}

Configure all aspects of Lightning for advanced usecases.

- [Level 14: Customize configs to run in production](/docs/pytorch/2.6.6/level-up/advanced/advanced_level_15.md) — Enable composable YAMLs

- [Level 15: Customize the trainer](/docs/pytorch/2.6.6/level-up/advanced/advanced_level_16.md) — Inject custom code into the trainer and modify the progress bar.

- [Level 16: Own the training loop](/docs/pytorch/2.6.6/level-up/advanced/advanced_level_17.md) — Learn all the ways of owning your raw PyTorch loops with Lightning.

- [Level 17: Enable advanced checkpointing](/docs/pytorch/2.6.6/level-up/advanced/advanced_level_18.md) — Enable composable or cloud based checkpoints.

- [Level 18: Master TPUs](/docs/pytorch/2.6.6/level-up/advanced/advanced_level_20.md) — Master TPUs and run on cloud TPUs.

- [Level 19: Train models with billions of parameters](/docs/pytorch/2.6.6/level-up/advanced/advanced_level_21.md) — Scale GPU training to models with billions of parameters

## Expert skills {#expert-skills}

Customize and extend Lightning for things like custom hardware or distributed strategies.

- [Level 21: Extend the Lightning CLI](/docs/pytorch/2.6.6/level-up/expert/expert_level_23.md) — Extend the functionality of the Lightning CLI.

- [Level 22: Integrate a custom cluster](/docs/pytorch/2.6.6/level-up/expert/expert_level_24.md) — Integrate a custom cluster into Lightning.

- [Level 23: Make your own profiler](/docs/pytorch/2.6.6/reference/tuning/profiler_expert.md) — Make your own profiler.

- [Level 24: Add a new accelerator or Strategy](/docs/pytorch/2.6.6/levels/expert_level_25.md) — Integrate a new accelerator or distributed strategy.

---

**Navigation**

- Parent: [Reference](/docs/pytorch/2.6.6/reference.md)