Search results
Welcome to KoboldAI on Google Colab, GPU Edition! KoboldAI is a powerful and easy way to use a variety of AI based text generation experiences. You can use it to write stories, blog posts, play...
This notebook allows you to download and use 4bit quantized models (GPTQ) on Google Colab.
This notebook provides an introduction to computing on a GPU in Colab. In this notebook you will connect to a GPU, and then run some basic TensorFlow operations on both the CPU and a GPU,...
31 lip 2024 · In this article, we will learn to use GPU i.e. Graphics Processing Unit in our google colab notebook. You can also refer to the video solution for this end which is attached at the end of this article.
Colab is a hosted service that lets you use Jupyter Notebooks with free access to GPUs and TPUs. It is well suited to machine learning, data science, and education.
GPUs and TPUs are different types of parallel processors Colab offers where: GPUs have to be able to fit the entire AI model in VRAM and if you're lucky you'll get a GPU with 16gb VRAM, even 3 billion parameters models can be 6-9 gigabytes in size. Most 6b models are even ~12+ gb.
ColabKobold | The (Unofficial) easy way to play KoboldAI in Google Colab (Works on phones) I finally managed to make this unofficial version work, its a limited version that only supports the GPT-Neo Horni model, but otherwise contains most features of the official version.
