we are ready for you! http://
sora.com (and check out how good the featured feed is getting)
MULTIMODAL AI
-
Sora AI Platform Launch: Creative Content Generation Ready
By
–
-
Predibase Wrapped 2024: AI Breakthroughs and 2025 Preview
By
–
Join us for "Predibase Wrapped: Our Greatest Hits of 2024"! We'll cover this year's major #AI breakthroughs and get a glimpse at 2025. Continued Training
VLM
Turbo LoRA
Multi-GPU Training
FP8 GPU Optimizations Dec. 19 @ 10am PT Register now: https://
pbase.ai/3VzLku4 -
Best-of-N Jailbreaking Techniques for Vision Language Models
By
–
Best-of-N isn't limited to text. We jailbroke vision language models by repeatedly generating images with different backgrounds and overlaid text in different fonts. For audio, we adjusted pitch, speed, and background noise. Some examples are here: https://
jplhughes.github.io/bon-jailbreaki
ng/#examples
… -
ChatLLM Teams: Interactive Data Exploration Across Multiple Modalities
By
–
Interactively explore, query, & summarize various data modalities (PDF/text, images, databases) with @AbacusAI ChatLLM Teams. This video shows how ChatLLM answers questions about any PDF! Try it out here: https://t.co/ATIFqtIsNK
— Kirk Borne (@KirkDBorne) 13 décembre 2024
————#LLMs #GenAI #AI #GenerativeAI pic.twitter.com/bVeqyolpGqInteractively explore, query, & summarize various data modalities (PDF/text, images, databases) with @AbacusAI ChatLLM Teams. This video shows how ChatLLM answers questions about any PDF! Try it out here: http://
chatllm.abacus.ai/?token=kirk
————
#LLMs #GenAI #AI #GenerativeAI -
Google’s Latest AI Innovations: NotebookLM and Gemini Deep Research
By
–
Until today, Audio Overviews in @GoogleAI NotebookLM was the most impressive AI innovation I'd seen this year. And then I tested @GoogleDeepMind
’s new @GeminiApp Deep Research. It's mind-blowing. -
ChatGPT sees my activities in real time: amazing result
By
–
I just let ChatGPT SEE everything I do, in real time. The result? MIND-BLOWING Curious to see that?
https://
youtu.be/Lv2a4Iy1zKk – I take it into my YouTube studio
– I take it into my LinkedIn creation
– I take it to play chess… #ChatGPT #AI #GPT4o -
DeepSeek-VL2: Enhanced Multimodal Model with Dynamic Tiling
By
–
Improvements from v1: > 2x high-quality training data vs DeepSeek-VL1
> Dynamic image tiling for flexible resolutions + efficient DeepSeek-MoE for LM
> 3-stage training + new multi-modal parallel strategies for efficiency -
DeepSeek-VL2 Achieves SoTA Vision Performance with Fewer Parameters
By
–
The whale strikes again! DeepSeekVL 2 > DeepSeek-VL2-Tiny, DeepSeek-VL2-Small, and DeepSeek-VL2, with 1.0B, 2.8B, and 4.5B activated parameters
> SoTA perf with similar or fewer activated parameters compared to Qwen 2 VL
> Excels at visual question answering, optical -
AI Tool Discovers Elephants and Monkeys Have Names
By
–
How cool is this: "Researchers have already leveraged the tool to discover that both African savannah elephants (Loxodonta africana) and common marmoset monkeys (Callithrix jacchus) have names." "The computer correctly matched [for elephants] the calls to the recipient 27.5% of