Featured tutorial
Convert a Hugging Face model to GGUF in Google Colab
Quantize and export models from the Hub to GGUF for llama.cpp and Ollama. Step-by-step workflow using Transformers, bitsandbytes, and Colab.
Read tutorial →GUIDE
Hugging Face & Local LLMs
Practical tutorials for Transformers, local models, embeddings, GGUF conversion, and open-source LLM workflows.
Ruslan Magana Vsevolodovna
Machine Learning Engineer · Data Scientist · Physicist
Genoa, Italy
Latest tutorials
How to download Hugging Face models to a custom cache folder
Organize model files, snapshots, and revisions with a custom HF_HOME directory and cache_dir.
Read tutorial →
Run FLAN-T5 and GPT locally with Gradio
Launch interactive demos for local models with an easy Gradio interface.
Read tutorial →
Fine-tune Llama 3 with Unsloth (LoRA + GGUF)
Fine-tune an instruction-following Llama 3 on a single Colab GPU with LoRA, then export to GGUF for local use.
Read tutorial →
Build a basic LLM (GPT) from scratch in Python
Implement a small GPT-style language model and understand the transformer loop end to end.
Read tutorial →
Fine-tune a transformer model for text classification
Prepare data, train, evaluate, and save your own transformer model with the Hugging Face Trainer.
Read tutorial →
Build embeddings and semantic search with Sentence Transformers
Create embeddings and a vector search pipeline with FAISS and Hugging Face Sentence Transformers.
Read tutorial →
Deploy a Hugging Face model with FastAPI
Build a production-ready API for inference with FastAPI and Uvicorn.
Read tutorial →
Use Hugging Face pipelines for rapid prototyping
Leverage pipelines for text, vision, speech, and more in just a few lines of code.
Read tutorial →FAQ
What is Hugging Face used for?
Hugging Face provides models, datasets, libraries, and tools for building and deploying machine learning and LLM workflows.
Can I run Hugging Face models locally?
Yes. Many models can be downloaded and run locally with Transformers, llama.cpp, Ollama, Gradio, or FastAPI depending on the format and hardware.
What is GGUF and why is it useful?
GGUF is a model format commonly used with llama.cpp and local inference tools. It helps run quantized models efficiently on local machines.