DALLE-pytorch

focused on amassing several of the bigger datasets of the time. Getting CC12M downloaded and trained on was something of an early milestone (robvanvolt's work). A lot of early work was like that though, shuffling through CC12M, COCO, etc. with the dalle-pytorch codebase until we got an avocado armchair. Christophe Schumann was an early contributor as well and great at organizing and rallying. He focused a lot on the early data scraping work for what would become the "LAION5B" dataset. I don't want to credit him with the coding and I'm ashamed to admit I can't recall who did much of the work there - but a distributed scraping program was developed (the name was something@home... not scraping@home?). I should also mention Romain Beaumont, an original member. His work on clip-retrieval was great and allowed us to eventually ship CLIP embeddings data for LAION5B and do semantic search over the dataset. He worked on a lot there however, including code for training and model architecture. Smart guy! I have heard he's now with Stability? Haven't confirmed though. The discord link on Phil Wang's readme at dalle-pytorch got a lot of traffic and a lot of people who wanted to pitch in with the scraping effort. Eventually a lot of people from Eleuther and many other teams mingled with us, some sort of non-profit org was created in Germany I believe for legal purposes. The dataset continued to grow and the group moved from training DALLE's to finetuning diffusion models. Eventually it was decided by core members that we should get a proper name and LAION was chosen. The `CompVis` team were great inspiration at the time and much of their work on VQGAN and then latent diffusion models basically kept us motivated. As I mentioned a personal motivation was Katherine Crowson's work on a variety of things like CLIP-guided vqgan, diffusion, etc. I believe Emad Mostaque showed up around the time GLIDE was coming out? I want to say he donated money for scrapers to be run on AWS to speed up data collection. I was largely hands off for much of the data scraping process and mostly enjoyed training new models on data we had. As with any online community things got pretty ill-defined, roles changed over, volunteers came/went, etc. I would hardly call this definitive and that's at least partially the reason it's hard to trace as an outsider. That much of the early history is scattered about GitHub issues and PR's can't have helped though.

g-transformers - "Discrete VAE", used as the backbone for OpenAI's DALL-E, reimplimented here (and other places) https://github.com/lucidrains/DALLE-pytorch - "VQVAE2" https://github.com/tgisaturday/dalle-lightning

similar scale to Open AI and then some because we are also working with another brilliant German team https://github.com/CompVis/ who has provided us with what they are calling a "VQGAN" (if you're not familiar) - which is a variational autoencoder for vision tokens with the neat trick from GAN-land of using a discriminator in order to produce fine details. https://github.com/CompVis/taming-transformers We use their pretrained VQGAN to convert an image into digits. We use another pretrained text tokenizer to convert words to digits. The digits both go into a Transformer architecture and a mask is applied to the image tokens in the transformer so that the text tokens can't see the image tokens. The digits come out and we encode them back into text and image respectively. Then, a perceptual loss is computed. Rinse, wash, repeat. Slowly but surely, text predicts image without ever having been able to actually _see_ the image. Insanity. Anyway, taking a caption and making a neural network output an image from it has again hit that "perfect intersection of machine learning, accessibility in education, and just plain cool". I don't know if you could fit it into the format of your YouTube channel but perhaps it would be a good match?