ERNIE-ViLG 2.0: Improving Text-to-Image Diffusion Model with Knowledge-Enhanced Mixture-of-Denoising-Experts
— AK (@_akhaliq) October 28, 2022
abs: https://t.co/TYNxPv2K5F
achieves a zero-shot FID-30k score of 6.75 on the MS-COCO dataset pic.twitter.com/VRUeIJRRVR
DiffusionDB: A Large-scale Prompt Gallery Dataset for Text-to-Image Generative Models
— AK (@_akhaliq) October 27, 2022
abs: https://t.co/nuhVrnsDw9
github: https://t.co/RTp4nNDYVR pic.twitter.com/OzzsrYNypo
The math of Gaussian Mixture Model Clustering can be tough for undergrads to grasp, but it gives a TON of insight into how GMM works!
— Chelsea Parlett-Pelleriti (@ChelseaParlett) October 26, 2022
I made this GMM math worksheet to do with my class. https://t.co/oAYyMvNSRE pic.twitter.com/6ApX4GBvj2
New Dreambooth method, train under 10 minutes without class images on multiple subjects
— AK (@_akhaliq) October 26, 2022
github: https://t.co/QO7UFwqXjf
colab: https://t.co/wIitnNdenM
reddit thread: https://t.co/gJlHikAt79 pic.twitter.com/7xKLtkzYza
Lafite2: Few-shot Text-to-Image Generation
— AK (@_akhaliq) October 26, 2022
abs: https://t.co/WBV6jmciIl pic.twitter.com/fdpejYhPet
Learn how Open Images V7 expands the Open Images dataset with new point-level label annotations, which provide localization information for 5.8k classes, and a new all-in-one visualization tool for better data exploration. https://t.co/1dbDuiifYT pic.twitter.com/oGGcTUJZpZ
— Google AI (@GoogleAI) October 25, 2022
Now that ML is so much about scaling up on all fronts, and so little about ideas, scientific awards (e.g. "NeurIPS oral") should be resource normalized.
— Rosanne Liu @ NeurIPS (@savvyRL) October 25, 2022
Student and independent researcher led research should NOT be evaluated the same way as industry-backed research. Period.
Stable Diffusion v1-5 @Gradio Demo is out on @huggingface Spaces by @runwayml
— AK (@_akhaliq) October 25, 2022
demo: https://t.co/iCMHIFKpGe pic.twitter.com/6vtBHyG0tC
big AI art news from Shutterstock. the company is:
— James Vincent (@jjvincent) October 25, 2022
- integrating OpenAI's DALL-E to their site to generate stock images
- launching a fund to reimburse Shutterstock creators whose images are sold to train AI art modelshttps://t.co/7OlhYYyPlB
Is it possible to transform sensitive user data in a way that 𝘨𝘶𝘢𝘳𝘢𝘯𝘵𝘦𝘦𝘴 fairness while still being useful for (unknown) downstream applications?
— DeepMind (@DeepMind) October 25, 2022
Introducing LASSI: a new provably fair representation learning method for high-dimensional data: https://t.co/ww1I6qd4jA pic.twitter.com/D4VpJLY1qL
Towards Real-Time Text2Video via CLIP-Guided, Pixel-Level Optimization
— AK (@_akhaliq) October 25, 2022
abs: https://t.co/qeHhRHxf4k
project page: https://t.co/ucoHXyVPAt pic.twitter.com/fYkMKXN5LY
Arcane Diffusion v3 - Updated dreambooth model now available on @huggingface
— AK (@_akhaliq) October 24, 2022
model: https://t.co/OiGSTfZnlv pic.twitter.com/78lTrv4t9J
