In an ELO comparison study of the same control types with approximately 150 participants, Stable Diffusion 3.5 Large ControlNets ranked first in user preference among similar models. (5/5)
MULTIMODAL AI
-
Stable Diffusion 3.5 Adds ControlNet Capabilities for Enhanced Control
By
–
We heard you want Control. So, today we are adding new capabilities to Stable Diffusion 3.5 Large with the release of three ControlNets: Blur, Canny, and Depth. The models are free for both commercial and non-commercial use under the permissive Stability AI Community
-
Google Gemini Adds Codebase Upload Feature
By
–
BREAKING : It seems like Google is releasing "codebase upload" for Gemini! It allows you to select a single folder per chat with up to 1000 files and 100MB. Seems like it is already rolling out (also correlates with a planned release on the 25 of November
-
OuteTTS v0.2: Compact Multilingual TTS with Voice Cloning
By
–
Smol TTS keeps getting better! Introducing OuteTTS v0.2 – 500M parameters, multilingual with voice cloning! 🔥
— Vaibhav (VB) Srivastav (@reach_vb) 25 novembre 2024
> Multilingual – English, Chinese, Korean & Japanese
> Cross platform inference w/ llama.cpp
> Zero-shot voice cloning
> Trained on 5 Billion audio tokens
> Qwen 2.5… pic.twitter.com/N4c4UkfrhzSmol TTS keeps getting better! Introducing OuteTTS v0.2 – 500M parameters, multilingual with voice cloning! > Multilingual – English, Chinese, Korean & Japanese
> Cross platform inference w/ llama.cpp
> Zero-shot voice cloning
> Trained on 5 Billion audio tokens
> Qwen 2.5 -
Qwen 2 VL + FLUX: Advanced Open Source Image Generation
By
–
Qwen 2 VL + FLUX > supports variation, img2img, inpainting, and controlnet-guided generation
> depth estimation and line detection for precise structural guidance
> aspect ratios up to 1536×1024 MIT licensed! -
AI Avatar Pickle Replaces You in Video Meetings
By
–
Se acabó eso de permanecer sentado frente a la cámara para tus reuniones en vídeo llamada. Con Pickle un avatar creado por IA lo hará por ti, así tú puedes moverte libremente.
— Juan Merodio (@juanmerodio) 25 novembre 2024
Impresionate 😲 pic.twitter.com/h4GGtgCIWkSe acabó eso de permanecer sentado frente a la cámara para tus reuniones en vídeo llamada. Con Pickle un avatar creado por IA lo hará por ti, así tú puedes moverte libremente. Impresionate
-
Event-Based Cameras: Technology Status and AI Applications
By
–
what about event based cameras? seemed dead forever still?
-
Major Open Source LLM Releases: Pixtral, Tülu Compete With Claude
By
–
Massive week for Open AI/ ML: @MistralAI Pixtral & Instruct Large – ~123B, 128K context, multilingual, json + function calling & open weights @allen_ai Tülu 70B & 8B – competive with claude 3.5 haiku, beats all major open models like llama 3.1 70B, qwen 2.5 and nemotron Llava
-
Apple Releases Fast CoreML Models for iPhone Performance
By
–
Pretty cool! @Apple released blazingly fast CoreML models AND an iOS app to run them on iPhone! ⚡
— Vaibhav (VB) Srivastav (@reach_vb) 23 novembre 2024
> S0 matches OpenAI's ViT-B/16 in zero-shot performance but is 4.8x faster and 2.8x smaller
> S2 outperforms SigLIP's ViT-B/16 in zero-shot accuracy, being 2.3x faster, 2.1x… pic.twitter.com/p9hPoajOtvPretty cool! @Apple released blazingly fast CoreML models AND an iOS app to run them on iPhone! > S0 matches OpenAI's ViT-B/16 in zero-shot performance but is 4.8x faster and 2.8x smaller > S2 outperforms SigLIP's ViT-B/16 in zero-shot accuracy, being 2.3x faster, 2.1x
-
Predibase Launches Visual Language Models Fine-tuning and Serving Beta
By
–
Big news: Predibase now supports fine-tuning & serving Visual Language Models (#VLMs) [beta]! Combine visual + text inputs for reasoning, captioning, VQA, multimodal gen, & image search. Fine-tune easily, deploy faster, and outperform OOTB models. Docs