A List of LLM resources to help get projects started
This list is focused on collecting the tools needed to get LLM-based projects started quickly including free resources, links to the most common APIs, examples, and template projects.
Contents
- 🔌 APIs
- 💻 Local Models
- 🎁 Offers and Promotions
- 🧰 Software and Tools
- 📚 Tutorials and Examples
- If you only have time for 8 minutes, let “Ryan Gosling” explain LLMs to you. YouTube
- “Getting started with huggingface.co in 15 min”. AssemblyAI, YouTube. This is what I used to run my first open-source model. Gives you some understanding of how to use the transformers Python library and what Models and Tasks are.
- “Neural Networks” (visualized). 3Blue1Brown (Grant Sanderson), YouTube, homepage: Gives you a nice intuition on how neural networks in general and transformers work in particular. I recommend the transformer/attention video (chapter 6) as it also gives a good sense of the scale of an LLM like GPT-3.
- “Neural networks from zero to hero” (i.e. from gradient descent to transformers), Andrej Karpathy, YouTube: step-by-step coding instructions that I recommend following and reproducing to get some technical understanding of what is happening in NN and LLMs
- Building an LLM Agent by Kevin Jablonka and other LLM posts
- OpenAI - Access to GPT models and other AI capabilities.
- Hugging Face - API access to a wide range of machine learning models. How to use Llama on HF
- Google Vertex AI - A suite of AI and machine learning APIs provided by Google Cloud. $150 credits upon signup
- Azure AI Studio - A collection of AI services and APIs offered by Microsoft Azure. Students start with $100 free Azure credits <== you will need an account here to use AI Studio
- Groq Exceptionally fast inference for a variety of models. Fairly high rate limits for free tier, though can't find exact documentation.
- Llama Run llama3 locally using
llm
package - Mixtral
- Phi
- Apple OpenELM, a family of Open Efficient Language Models. 270M, 450M, 1.1B and 3B parameters
- LLM CLI Utility from Datasette
- huggingface.co, The GitHub for models. Use and share your NN models.
- LangChain, LangChain is a framework for developing applications powered by LLMs
- axolotl, if you want to get into fine-tuning
- on new Macs, mlx is a very fast way to run LLMs
- access to LLMs via API via Groq (limited to 30 req/min)
- 100 req/day at Awan LLM
- Awesome-LLM, a very deep awesome list with all the relevant papers and projects.
Contributions of any kind welcome, just follow the guidelines!
Thanks goes to these contributors!
Thank you also to Sterling Baird, who started collecting many of the software links included here.