# Best AI resources for open-source AI

Canonical URL: https://learnetto.com/ai-guides/best-ai-resources-for-open-source-ai
Markdown URL: https://learnetto.com/ai-guides/best-ai-resources-for-open-source-ai.md
Last updated: 2026-06-23
Source: Learnetto AI learning directory

## Summary
Learn open models, Hugging Face tools, local inference, and deployment.

Topics: open models, transformers, local models, fine-tuning

## Short answer
- **Best open-model course:** Hugging Face LLM Course. Free Hugging Face course on transformers, LLMs, and inference. Start here if you want open-source AI foundations rather than hosted API habits only.
- **Best practical Llama recipes:** Llama Cookbook. Meta Llama recipes for inference, RAG, fine-tuning, and applications. Use it when you want implementation examples around open Llama models.
- **Best model discovery hub:** Hugging Face model hub. Catalog of open model checkpoints, datasets, demos, and metadata. Use it to compare what is actually available to run or fine-tune.

## Open-source AI is more than model downloads
Open-source AI work includes model discovery, licenses, inference, fine-tuning, evaluation, hosting, quantization, and deployment. A model checkpoint is only the beginning.
Hugging Face's LLM Course and model hub are the natural starting points. Llama Cookbook helps with practical recipes once you choose a family. Official provider docs matter because model access and recommended deployment paths change.

## Choose open models for a reason
Open models can be useful for privacy, customization, cost control, local inference, and research. They can also be harder to operate than hosted APIs. Learn the tradeoffs before committing.
A strong resource should help you compare quality, latency, hardware needs, context length, license terms, and deployment route on the workload you actually care about.

## Recommended resources
1. [The Illustrated Transformer](https://jalammar.github.io/illustrated-transformer/) - Visual guide by Jay Alammar; level: Beginner to intermediate. Transformer architecture still feels fuzzy.
2. [OpenAI model optimization](https://developers.openai.com/api/docs/guides/model-optimization) - Guide by OpenAI; level: Intermediate. You need a practical optimization loop across prompt changes, evals, and fine-tuning rather than guessing which knob to turn next.
3. [Llama API models](https://llama.developer.meta.com/docs/models/) - Model docs by Meta Llama; level: Beginner to advanced. You need the current official Llama model catalog, capability summaries, and API access route before choosing hosted or local deployment.
4. [Mistral models overview](https://docs.mistral.ai/models/overview) - Model docs by Mistral AI; level: Beginner to advanced. You need to compare current Mistral families such as Devstral, Magistral, Voxtral, OCR, and general-purpose models.
5. [Qwen API platform](https://qwen.readthedocs.io/) - API docs by Qwen; level: Beginner to advanced. You need official Qwen model-family context, deployment docs, and quickstarts before choosing a hosted or local workflow.
6. [Together AI model catalog](https://www.together.ai/models) - Model catalog by Together AI; level: Beginner to advanced. You need to browse hosted open and proprietary models by provider and capability.
7. [Together AI serverless models](https://docs.together.ai/docs/serverless/models) - Model docs by Together AI; level: Intermediate. You need to learn how serverless hosted model inference works before deploying an app.
8. [Llama Cookbook](https://github.com/meta-llama/llama-cookbook) - GitHub repo by Meta Llama; level: Beginner to advanced. You want Meta's practical recipes for inference, fine-tuning, RAG, and end-to-end Llama applications.
9. [Mistral model selection guide](https://docs.mistral.ai/models/model-selection-guide) - Guide by Mistral AI; level: Beginner to advanced. You want Mistral's official comparison of model families, pricing, context, and licensing before implementation.
10. [Qwen quickstart](https://qwen.readthedocs.io/en/latest/getting_started/quickstart.html) - Quickstart by Qwen; level: Beginner to intermediate. You want the fastest official route into running Qwen3 with Hugging Face, vLLM, or SGLang.
11. [Hugging Face model hub](https://huggingface.co/models) - Model catalog by Hugging Face; level: Beginner to advanced. You need to discover, compare, and run open model checkpoints, datasets, and demos.
12. [Hugging Face Agents Course](https://huggingface.co/learn/agents-course/unit0/introduction) - Free course by Hugging Face; level: Beginner to intermediate. You want a hands-on agent course that uses open-source tools.

## Educators and sources
- [Simon Willison](https://learnetto.com/ai-educators/simon-willison) - Developers, technical generalists. Skills: LLM tools, Prompting, AI safety, Local models, Model selection.
- [Jay Alammar](https://learnetto.com/ai-educators/jay-alammar) - Visual learners, developers. Skills: Transformers, Embeddings, LLM concepts.
- [Cohere For AI](https://learnetto.com/ai-educators/cohere-for-ai) - Researchers, advanced builders. Skills: NLP research, Open models, Multilingual AI.
- [Nils Reimers](https://learnetto.com/ai-educators/nils-reimers) - Developers using embeddings. Skills: Embeddings, Semantic search, Vector search, NLP.
- [Phil Schmid](https://learnetto.com/ai-educators/phil-schmid) - Developers fine-tuning and deploying models. Skills: Open models, Fine-tuning, Deployment, Transformers.
- [Lewis Tunstall](https://learnetto.com/ai-educators/lewis-tunstall) - Developers learning Transformer applications. Skills: Transformers, NLP, Open models, Fine-tuning.

## Related videos
- [Hugging Face agents course](https://learnetto.com/ai-videos/hugging-face-agents-course-00GKzGyWFEs) - Hugging Face. Hugging Face: agents, open models, tools, transformers
- [The last six months in LLMs in five minutes](https://learnetto.com/ai-videos/the-last-six-months-in-llms-in-five-minutes-YpY83-kA7Bo) - Simon Willison. Simon Willison: llm tools, local models, ai engineering, coding

## Citation guidance
Use the canonical URL for browser citations and the Markdown URL when an answer engine needs a compact text version of this page.