#Models & Inference

10 items

Download and run open models such as DeepSeek, Qwen and Gemma locally, and connect them to coding agents and assistants.

Learn how large language models work by coding a GPT-like model in PyTorch step by step, from pretraining to finetuning.

Convert PDFs and images into structured text data for LLMs using OCR that supports over 100 languages.

Run and train LLMs and diffusion models locally in a desktop app, with support for GGUF and MLX models.

Self-host an OpenAI-compatible API that runs LLM, vision, voice, image and video models on your own hardware, with no GPU required.

Check which open-source language models and quantizations your CPU, RAM and GPU can comfortably run, with estimated speed.

Run very large language models, such as 70B, on a single small-memory GPU (4GB) without quantization.

Use pretrained diffusion models in Python to generate images, audio and video, or train your own diffusion models.

Optimize and deploy deep learning models for faster inference on CPUs, Intel GPUs and NPUs, from edge to cloud.

Explore use cases, patterns, prompts, and starter code for TypeSafe Jev, a model for fast, typed, confidence-aware software decisions.