Tan Minh Dinh @tanmdinh

VideoJAM is our new framework for improved motion generation from @AIatMeta We show that video generators struggle with motion because the training objective favors appearance over dynamics. VideoJAM directly adresses this **without any extra data or scaling** 👇🧵

Finally, a method that successfully personalizes me without compromising the model's prior! On the right, is the result of Nested Attention, which clearly outperforms other methods by adhering to "pointillism" while still capturing my likeness—confirmed by my dear wife!

*Shortcut models* are a plug-and-play replacement for diffusion models that can generate in a single step (or more). This speeds up inference by up to 128x. Shortcut models are trained end-to-end, and do not require a separate distillation phase or learning schedules.

Do we still need codebook/quantization for scalable autoregressive visual generation? No! Thrilled to share our latest work on scaling w/ continuous tokens. We observe power-law scaling behavior on val loss, and obtain SOTA coco FID and GenEval score. arxiv.org/abs/2410.13863

keyboard_arrow_left Previous keyboard_arrow_right Next

🎥 Today we’re premiering Meta Movie Gen: the most advanced media foundation models to-date. Developed by AI research teams at Meta, Movie Gen delivers state-of-the-art results across a range of capabilities. We’re excited for the potential of this line of research to usher in…

Gemini 1.5 Model Family: Technical Report updates now published In the report we present the latest models of the Gemini family – Gemini 1.5 Pro and Gemini 1.5 Flash, two highly compute-efficient multimodal models capable of recalling and reasoning over fine-grained information…

keyboard_arrow_left Previous keyboard_arrow_right Next

Here’s an early preview of ElevenLabs Music. All of the songs in this thread were generated from a single text prompt with no edits. Title: It Started to Sing Style: “Pop pop-rock, country, top charts song.”

# explaining llm.c in layman terms Training Large Language Models (LLMs), like ChatGPT, involves a large amount of code and complexity. For example, a typical LLM training project might use the PyTorch deep learning library. PyTorch is quite complex because it implements a very…

⚡️SD3-Turbo: Fast High-Resolution Image Synthesis with Latent Adversarial Diffusion Distillation Following Stable Diffusion 3, my ex-colleagues have published a preprint on SD3 distillation using 4-step, while maintaining quality. The new method – Latent Adversarial Diffusion…

keyboard_arrow_left Previous keyboard_arrow_right Next

Snap presents MyVLM Personalizing VLMs for User-Specific Queries Recent large-scale vision-language models (VLMs) have demonstrated remarkable capabilities in understanding and generating textual descriptions for visual content. However, these models lack an understanding of

Stabillity AI presents Stable Diffusion 3 Scaling Rectified Flow Transformers for High-Resolution Image Synthesis Diffusion models create data from noise by inverting the forward paths of data towards noise and have emerged as a powerful generative modeling technique for…

Sora: A Review on Background, Technology, Limitations, and Opportunities of Large Vision Models Sora is a text-to-video generative AI model, released by OpenAI in February 2024. The model is trained to generate videos of realistic or imaginative scenes from text instructions and…

I am excited and honored to co-organize #SyntaGen - a #CVPR2024 #workshop on Harnessing Generative Models for Synthetic Visual Dataset. We are calling for cutting-edge paper submissions. Detail information can find below or in our website syntagen.github.io

Announcing FlashAttention-2! We released FlashAttention a year ago, making attn 2-4 faster and is now widely used in most LLM libraries. Recently I’ve been working on the next version: 2x faster than v1, 5-9x vs standard attn, reaching 225 TFLOPs/s training speed on A100. 1/

keyboard_arrow_left Previous keyboard_arrow_right Next

Oops haven't tweeted too much recently; I'm mostly watching with interest the open source LLM ecosystem experiencing early signs of a cambrian explosion. Roughly speaking the story as of now: 1. Pretraining LLM base models remains very expensive. Think: supercomputer + months.…

Scaling up GANs for Text-to-Image Synthesis present our 1B-parameter GigaGAN, achieving lower FID than Stable Diffusion v1.5, DALL·E 2, and Parti-750M. It generates 512px outputs at 0.13s, orders of magnitude faster than diffusion and autoregressive mingukkang.github.io/GigaGAN/…

Stable Diffusion generates beautiful images, but can it be used for open-world recognition? Try Demo! huggingface.co/spaces/xvjiaru… Our #CVPR2023 paper shows that the pre-trained diffusion model indeed is a good image parser, allows for open-vocabulary segmentation and detection.

MEGANE: Morphable Eyeglass and Avatar Network abs: arxiv.org/abs/2302.04868 project page: junxuan-li.github.io/megane/

3DAvatarGAN: Bridging Domains for Personalized Editable Avatars abs: arxiv.org/abs/2301.02700 project page: rameenabdal.github.io/3DAvatarGAN/

This looks great! Simple idea, nice clean code implementation... Takes a key idea from diffusion (iterative improvement over multiple steps) but with a new (far less mathy) approach. Well done to the authors :)

AK @_akhaliq

3 years ago

This looks great! Simple idea, nice clean code implementation... Takes a key idea from diffusion (iterative improvement over multiple steps) but with a new (far less mathy) approach. Well done to the authors :)

5 75 332 0 101

Download Image