Like any tool only as good as the user ! We crafted these weird and wonderful images using OpenAI's Dalle-2
#dalle2 #dalle #GenerativeAI #ai #ArtificialIntelligence #MachineLearning #deeplearning https://
buff.ly/3MB7lEI
MULTIMODAL AI
-
Creative AI Images Generated with DALL-E 2
By
–
-
High-Resolution Video Synthesis Using Latent Diffusion Models
By
–
Align your Latents: High-Resolution Synthesis with Latent Diffusion Models https://
bit.ly/41QLfCO #AI #DeepLearning #MachienLearning #DataScience #GenAI -
Abacus AI Vector Matching Engine for Deep Learning Similarity Search
By
–
QBE (Query-By-Example) does feature vector similarity search. Well, here's something better… @abacusai has built & deployed a Vector Matching Engine to search vector embeddings of images, documents, and more, using #DeepLearning models: https://
abacus.ai/vectormatching
*
#AI #ML #BigData -
Talk to ChatGPT with microphone for voice responses
By
–
4. Talk to ChatGPT: Engage in conversations with ChatGPT using your microphone and receive voice responses in return. Find it at http://
chrome.google.com. -
Style Transfer: Apply Artistic Painting Styles to Any Image
By
–
10. Style transfer Given a painting of your favorite artist, and any other random picture, the application should output the same picture using the style from the painting. This one is not easy, but the results are pretty cool.
-
VideoChat: Integrating Language and Video Models for Understanding
By
–
Meet VideoChat: Integrating Language and Models to Boost Understanding
-
MidJourney Generates Defense Intelligence Agency Imagery
By
–
Typed the following prompt into MidJourney: "Defense intelligence agency Photorealistic, cinematic –ar 2:1" Result was pretty imaginative. Not sure what all that stuff in the courtyard is, but looks pretty high tech. @DefenseIntel #intelligence
-
MultiModal-GPT: Vision Language Model for Multi-Round Dialogue
By
–
10/ MultiModal-GPT – a vision and language model for multi-round dialogue with humans; the model is fine-tuned from OpenFlamingo, with LoRA added in the cross-attention and self-attention parts of the language model.
-
InstructBLIP Achieves State-of-the-Art Zero-Shot Visual-Language Performance
By
–
6/ InstructBLIP – explores visual-language instruction tuning based on the pre-trained BLIP-2 models; achieves state-of-the-art zero-shot performance on 13 held-out datasets, outperforming BLIP-2 and Flamingo.