(2/12) PaLM-E: An Embodied Multimodal Language Model
Authors: @DannyDriess
, @xf1280
, Mehdi S. M. Sajjadi, @coreylynch
, @achowdhery
, @brian_ichter
, @ayzwah
, @JonathanTompson
, @QuanVng
, @TianheYu
, @wenlong_huang
, @YevgenChebotar
, @psermanet
, @duck et. al.
MULTIMODAL AI
-
PaLM-E: Embodied Multimodal Language Model for Robotics
By
–
-
New Transformer-based Image Segmentation Model with Zero-shot Capabilities
By
–
A new image segmentation model that can segment almost anything via prompt.
— Jean de Dieu Nyandwi (@Jeande_d) 5 avril 2023
– Zero-shot generalization
– Based on Transformers
– Code and dataset released
– 632M + 4M params
– Can be prompted via background points, mask, and bounding box. https://t.co/etBlCd9yCjA new image segmentation model that can segment almost anything via prompt. – Zero-shot generalization
– Based on Transformers
– Code and dataset released
– 632M + 4M params
– Can be prompted via background points, mask, and bounding box. -
Niji-Journey V5 Anime-Focused AI Image Generation Tool Launch
By
–
Come play with Niji-Journey V5 this week, it's an anime-focused version of Midjourney (type /settings and click "Niji Version 5" on our bots). We also have a dedicated server for finding like-minded friends and native-language chat https://
discord.gg/nijijourney -
SMS vs Phone Calls: Choosing the Right Interface for AI
By
–
Went with SMS so I don't have to be by a computer, but maybe phone call via Whisper would feel more magical pic.twitter.com/ObDIlBOt4N
— Yohei (@yoheinakajima) 4 avril 2023Went with SMS so I don't have to be by a computer, but maybe phone call via Whisper would feel more magical
-
ControlNet: Enhanced Control for Image Generation with Stable Diffusion
By
–
🚀Blog-tastic Tuesdays🚀
— Satya Mallick (@LearnOpenCV) 4 avril 2023
ControlNet gives us more control over the image generation process when using the Img2Img method.https://t.co/VjiWZvG7wB
Our latest article discusses the architecture, training process, and different techniques for the ControlNet Stable Diffusion model.… pic.twitter.com/bNzYJESG3BBlog-tastic Tuesdays
ControlNet gives us more control over the image generation process when using the Img2Img method. https://
learnopencv.com/controlnet/ Our latest article discusses the architecture, training process, and different techniques for the ControlNet Stable Diffusion model. -
Autonomous Task Planning from Camera Analysis Systems
By
–
There’s lots of research here so definitely possible. You’d basically have a listener looking at the analyzed output of cameras etc, and it would create tasks based on a core objective which it can follow.
-
Midjourney launches /Describe, live test with ChatGPT and GPT4
By
–
#Midjourney just released the /Describe feature and we are testing it live → https://youtu.be/aAobYoM756s The possibilities are really huge #Midjourney5 #Midjourneyv5 #ChatGPT #GPT4
-
Rapid AI Image Generation Speed Amazes Users
By
–
How do you generate these beautiful pictures so fast? He posted this only 16 minutes before!
-
New /describe Command Transforms Images into Words
By
–
Today we're releasing a /describe command that lets you transform images-into-words. Give it a shot! We think this tool will transform your liguistic-visual process both in terms of creative power and discovery.
-
HuggingGPT combines ChatGPT and AI models
By
–
3. HuggingGPT: Solving AI Tasks with ChatGPT and its Friends in HuggingFace HuggingGPT is a system that combines ChatGPT with a variety of AI models to tackle complex tasks across language, vision, and speech for more versatile artificial intelligence. https://arxiv.org/abs/2303.17580