Just a couple weeks later, Stable Diffusion V2.1 was released, which was trained for longer on more data!
MULTIMODAL AI
-
Stable Diffusion 2.0 Released with Inpainting and Super-Resolution
By
–
We followed that up shortly with the release of Stable Diffusion 2.0, which also came with a suite of models for inpaintining. depth-to-image, and super-resolution!
-
Stable Diffusion and DreamStudio AI officially released August 2022
By
–
In Aug 2022, @StableDiffusion (in collaboration with Runway and CompVis research group) and @DreamStudioAI were both officially released!
-
AutoDecoding Latent 3D Diffusion Models for Asset Generation
By
–
AutoDecoding Latent 3D Diffusion Models
— AK (@_akhaliq) 12 juillet 2023
paper page: https://t.co/ftMUZW1hxg
present a novel approach to the generation of static and articulated 3D assets that has a 3D autodecoder at its core. The 3D autodecoder framework embeds properties learned from the target dataset in… pic.twitter.com/IPja1uHPRNAutoDecoding Latent 3D Diffusion Models paper page: https://
huggingface.co/papers/2307.05
445
… present a novel approach to the generation of static and articulated 3D assets that has a 3D autodecoder at its core. The 3D autodecoder framework embeds properties learned from the target dataset in -
Collaborative Score Distillation for Consistent Visual Synthesis
By
–
Collaborative Score Distillation for Consistent Visual Synthesis
— AK (@_akhaliq) 12 juillet 2023
paper page: https://t.co/Lb1yiocpqq
Generative priors of large-scale text-to-image diffusion models enable a wide range of new generation and editing applications on diverse visual modalities. However, when… pic.twitter.com/NzO05zCI8fCollaborative Score Distillation for Consistent Visual Synthesis paper page: https://
huggingface.co/papers/2307.04
787
… Generative priors of large-scale text-to-image diffusion models enable a wide range of new generation and editing applications on diverse visual modalities. However, when -
Emu: Multimodal Foundation Model for Image and Text Generation
By
–
Generative Pretraining in Multimodality paper page: https://
huggingface.co/papers/2307.05
222
… present Emu, a Transformer-based multimodal foundation model, which can seamlessly generate images and texts in multimodal context. This omnivore model can take in any single-modality or multimodal data -
Pika Labs Releases Image-Conditioned Video Generation Feature
By
–
Pika labs releases image-conditioned video generation, upload an image, and the model will animate the image, prompt is “a girl in the wind”pic.twitter.com/QVrYYJjxF6
— AK (@_akhaliq) 12 juillet 2023Pika labs releases image-conditioned video generation, upload an image, and the model will animate the image, prompt is “a girl in the wind”
-
One-2-3-45: Convert Any Single Image to 3D Mesh in 45 Seconds
By
–
One-2-3-45: Any Single Image to 3D Mesh in 45 Seconds without Per-Shape Optimization @Gradio demo is out
— AK (@_akhaliq) 11 juillet 2023
demo: https://t.co/IlfEzhOYPVpic.twitter.com/KgbbfnwRi1One-2-3-45: Any Single Image to 3D Mesh in 45 Seconds without Per-Shape Optimization @Gradio demo is out demo: https://
huggingface.co/spaces/One-2-3
-45/One-2-3-45
… -
Reenactment Network with ReshotAI Key Points
By
–
Reenactment network with reshotAI key points by @alexcarliera
— AK (@_akhaliq) 11 juillet 2023
pic.twitter.com/UtRfnrP2w9Reenactment network with reshotAI key points by @alexcarliera
-
Objaverse-XL: Massive 10M+ 3D Objects Dataset Released
By
–
Objaverse-XL: A Universe of 10M+ 3D Objects
— AK (@_akhaliq) 11 juillet 2023
paper: https://t.co/Q1iVTAg29Q
Natural language processing and 2D vision models have attained remarkable proficiency on many tasks primarily by escalating the scale of training data. However, 3D vision tasks have not seen the same… pic.twitter.com/ptBAwG9bNFObjaverse-XL: A Universe of 10M+ 3D Objects paper: https://
objaverse.allenai.org/objaverse-xl-p
aper.pdf
… Natural language processing and 2D vision models have attained remarkable proficiency on many tasks primarily by escalating the scale of training data. However, 3D vision tasks have not seen the same