KerasCV now includes a Segment Anything implementation, which enables you to do image segmentation using only "prompt points". No training data needed. With the JAX backend, it runs ~5x faster on GPU than the original PyTorch implementation. Told you JAX was fast. Guide:
MULTIMODAL AI
-
ChatGPT Struggles with Negation Prompts in Image Generation
By
–
Anyone notice ChatGPT seems to underperform on negation prompts?
— Allie K. Miller (@alliekmiller) 4 janvier 2024
One viral example: the man asking Dall-E to remove the chopsticks from the image of the ramen, but they just keep moving around the bowl. https://t.co/w6anLdV8hiAnyone notice ChatGPT seems to underperform on negation prompts? One viral example: the man asking Dall-E to remove the chopsticks from the image of the ramen, but they just keep moving around the bowl.
-
Single Generation Produces 80-Frame 1280x Video with Remarkable Consistency
By
–
Dang!
— fofr (@fofrAI) 4 janvier 2024
80 frames
a single generation
10 second video at 1280x
Remarkable consistency here over the time frame (10 minutes to generate though) https://t.co/Dgl6FfdYmU pic.twitter.com/NDm8rqeWeeDang! 80 frames
a single generation
10 second video at 1280x Remarkable consistency here over the time frame (10 minutes to generate though) -
i2vgen Model Experiment: Character Movement Consistency Across Prompts
By
–
An experiment with i2vgen model – I keep the seed and params the same but vary prompt by tweaking action (standing, dancing, crying, etc)
— fofr (@fofrAI) 4 janvier 2024
Some pretty solid character movement, consistency is pretty good, only the occasional scary distortion.
12 variations. https://t.co/Dgl6FfdYmU pic.twitter.com/rqTeLC8jt9An experiment with i2vgen model – I keep the seed and params the same but vary prompt by tweaking action (standing, dancing, crying, etc) Some pretty solid character movement, consistency is pretty good, only the occasional scary distortion. 12 variations.
-
Image Cropping Affects Model Scale Recognition Outcomes
By
–
For context, this has to do with the size of the image and the cropping, the model can read the scale but it's getting cut off, you can do a simple re-size to see it change the outcome.
-
Generate Custom AI Character Images with Detailed Prompts
By
–
Now, just wait a few minutes and you can write any prompt to see your character, Emily, in different settings and situations. Emily on a horse, clad in armor within a magical forest, or even transforming her into an action figure.
-
EverArt AI Enables Pose Transfer for Character Image Generation
By
–
But it doesn't stop there. @everartai also supports pose transfer. Found a pose you love? Simply transfer it to your character. Here's Emily striking a pose in a paparazzi shot. Left is the original photo, right is Emily.
-
Creating AI Influencers: Meet Emily on EverArt
By
–
AI influencers are set to be a major trend in 2024. @everart is the best place to create a consistent AI character. Just a few photos – real or AI-generated – and you can create a unique AI influencer. Let's see how to do it!
Meet Emily! -
LangChain Breast Ultrasound Report Generation Model
By
–
New Paper Alert: Breast Ultrasound Report Generation using LangChain This recent paper out of Korea Advanced Institute of Science and Technology uses LangChain and classical image classification models to help with medical report generation https://
arxiv.org/pdf/2312.03013
.pdf
… -
ChatGPT UI/UX Efficiency and AI Progress with Jerry Liu
By
–
ChatGPT enhances UI/UX efficiency by 30% by simplifying AI interaction, transforming coding and creative tasks. Jerry Liu's talk demystifies AI's progress, exploring multimodal models, RAG systems, and their impact on data interaction: https://
youtu.be/kZxl4gpe3OM