Image: Stable diffusion, Imagen, DALL-E 2 ◦ Audio: Riffusion • AI Tools for Everyday Use ◦ Productivity ◦ Business ◦ Design, and more ◦ A curated list of 600+ tools! • Q&A Plan of the day North America Connect
enables computers to mimic human intelligence & behaviour A subset of AI that includes statistical techniques to learn from experience and make predictions or decisions A subset of ML exposing multilayered neural networks to vast amount of data to make decisions Create new written, visual, and auditory content given prompts or existing data
Generative models are used to generate new outputs from scratch. So, given a prompt, it can imagine the rest. In the case of GPT, it can only generate text.
pretrained on *LOTS of data. Think ~all of the publically available text on the internet. Models trained on LOTS of data are called Large Language Model (LLM). ChatGPT
specific architecture of neural networks. Transformers were first published in 2017 by Google researchers. It allowed researchers to train models efficiently on huge datasets. https://arxiv.org/pdf/1706.03762.pdf ChatGPT
Bing Chat Bard (PaLM 2) ChatGPT (CPT-3.5) Bing Chat (GPT-4) Real-time access to internet for free ✅ ✅ Voice input ✅ Export generated text / table ✅ Provide multiple drafts for the response ✅ (re-generate) Cross-app connectivity (Search, Maps) ✅ Summarize web-pages for free ✅ ✅ Good at coding ✅ Good at Math / logic ✅ Faster ✅ More Creative and Coherent ✅
developed by Stability AI • Uses “diffusion” to generate high-quality images from text • The diffusion process involves iteratively updating a set of image pixels based on a diffusion equation • Trained on a large dataset of images and text • The model learns to create images by comparing its output to the ground truth images in the dataset
text-to-image diffusion model capable of generating spectrogram • The spectrogram can then be translated into sound • It uses Torchaudio to read the frequency and time to play the sound • Users can create waveforms, visualize them, listen to what they sound like, and download the audio clips
Uses OpenAI GPT-3 to generate human-like text and is trained on publicly available code from GitHub. It is available in Visual Studio Code, JetBrains, and Neovim. Similar to DuetAI by Google