Why Leonardo over MJ? I haven’t used it in a long while, should check it out again
MULTIMODAL AI
-
Gemini Team at NeurIPS 2023 Conference Booth
By
–
On my way to @NeurIPSConf #NeurIPS2023 where I'm presenting 0 papers 🙃
— Oriol Vinyals (@OriolVinyalsML) 10 décembre 2023
Come chat with the Gemini ♊️ team at @GoogleDeepMind / @GoogleResearch booths on Tue/Wed from 1:30 to 3:00 PM (NOLA time). Quite a few Gemini team members will be there! pic.twitter.com/d8uYdMhiKROn my way to @NeurIPSConf #NeurIPS2023 where I'm presenting 0 papers Come chat with the Gemini team at @GoogleDeepMind / @GoogleResearch booths on Tue/Wed from 1:30 to 3:00 PM (NOLA time). Quite a few Gemini team members will be there!
-
LLMs Understanding Time and Place with OS Integration
By
–
Can LLMs understand time and place?
— Aaron Ng (@localghost) 10 décembre 2023
Here’s @apolloaiapp reading my location and calendar availability to propose a calendar event. A hint of what LLMs look like once they start working with your OS. pic.twitter.com/n4K3rTzXvCCan LLMs understand time and place? Here’s @apolloaiapp reading my location and calendar availability to propose a calendar event. A hint of what LLMs look like once they start working with your OS.
-
DALL-E 3 Image Captioner Improves Recreation with Synthetic Data
By
–
@OpenAI DALL·E 3's specialized image captioner creates precise, informative descriptions, enhancing image recreation over its predecessor by using improved synthetic data.
-
DALL·E 3 Advances Text-to-Image Generation Despite Spatial Limitations
By
–
Despite challenges in spatial awareness and accuracy, DALL·E 3's detailed text-to-image outputs signal a major advance, excelling in compositional benchmarks. Learn more: https://
youtu.be/Ilu4Nyb5_As -
SDXL Fine-Tune Model Sees Faces in Everything
By
–
Pareidolia SDXL fine-tune sees faces in everything.
-
CLIP Binding Problem and CAB Instance Discussion
By
–
Thank you for the reference! The original motivation was indeed to investigate the binding problem of CLIP, and in the paper we discuss that CAB is an instance of the binding problem.
-
Prompt Variation Effects on Similarity Scores in Image Retrieval
By
–
Haven't done image retrieval experiments. We did vary prompts and observed things like the similarity score for "lemon is eggplant" is higher than "lemon is purple" etc.
-
Zero-shot Classification with Color Labels and Part-whole Attributes
By
–
Yes, we did zero-shot classification / candidate matching and labels we used for the fruit-vegetable dataset are “red”,
“yellow”, “purple”, “green”, and “orange”. We also observed CAB for part-whole attributes, although it was not as severe as the color attributes. -
Attention Modulation Improves Object Localization in Autoregressive Models
By
–
Thanks! Attention modulation (like you did) is I think why autoregressive loss based models can localize objects better. Nice work too!