ByteDance presents INFP — an AI that turns any portrait and audio file into a talking or even singing head! 🎤🎥
Key Features of INFP:
Realistic facial expressions and rhythmic head movements based on audio input.
The standout feature: Dynamic switching between speaking and listening states, guided by the input sound.
While there are other neural networks out there, INFP’s ability to react to dynamic audio is what sets it apart. 🔊👂
Soon, AI agents like these will be chatting with you in real time, possibly even replacing customer support agents, offering more personalized interactions.
Check out the project page on GitHub for more!
#AI #INFP #ByteDance #VoiceGeneration
Brain Storm
Key Features of INFP:
Realistic facial expressions and rhythmic head movements based on audio input.
The standout feature: Dynamic switching between speaking and listening states, guided by the input sound.
While there are other neural networks out there, INFP’s ability to react to dynamic audio is what sets it apart. 🔊👂
Soon, AI agents like these will be chatting with you in real time, possibly even replacing customer support agents, offering more personalized interactions.
Check out the project page on GitHub for more!
#AI #INFP #ByteDance #VoiceGeneration
Brain Storm
This media is not supported in your browser
VIEW IN TELEGRAM
⛽ VoRA: Vision as LoRA ⛽
👉#ByteDance unveils Vision as LoRA (VoRA), a novel paradigm converting LLMs into Multimodal Large Language Models (MLLMs) by integrating vision-specific LoRA layers. LoRA -> LLM. All training data, codes, and model weights available💙
👉Review https://t.ly/guNVN
👉Paper https://arxiv.org/pdf/2503.20680
👉Repo https://github.com/Hon-Wong/VoRA
👉Project georgeluimmortal.github.io/vora-homepage.github.io/
👉#ByteDance unveils Vision as LoRA (VoRA), a novel paradigm converting LLMs into Multimodal Large Language Models (MLLMs) by integrating vision-specific LoRA layers. LoRA -> LLM. All training data, codes, and model weights available💙
👉Review https://t.ly/guNVN
👉Paper https://arxiv.org/pdf/2503.20680
👉Repo https://github.com/Hon-Wong/VoRA
👉Project georgeluimmortal.github.io/vora-homepage.github.io/
ezgif-8120c4563e81c3.mp4
510.6 KB
🥶 OmniHuman-1.5 🥶
👉#ByteDance proposes a novel framework designed to generate character animations that are not only physically plausible but also semantically coherent and expressive. Coherency with speech's rhythm, prosody and semantic content. Impressive results but no code 🥺
👉Review https://t.ly/CnRmX
👉Paper arxiv.org/pdf/2508.19209
👉Project omnihuman-lab.github.io/v1_5/
👉Repo 🥺
👉#ByteDance proposes a novel framework designed to generate character animations that are not only physically plausible but also semantically coherent and expressive. Coherency with speech's rhythm, prosody and semantic content. Impressive results but no code 🥺
👉Review https://t.ly/CnRmX
👉Paper arxiv.org/pdf/2508.19209
👉Project omnihuman-lab.github.io/v1_5/
👉Repo 🥺
This media is not supported in your browser
VIEW IN TELEGRAM
🐙Human-Centric Video Generation🐙
👉Tsinghua & #ByteDance unveil HuMo: a unified, human-centric video generation framework designed to produce HQ fine-grained, and controllable human videos from multimodal inputs. It supports strong text prompt following, consistent subject preservation, synchronized audio-driven motion. Repo released under Apache2.0💙
👉Review https://t.ly/3S8Yb
👉Paper https://arxiv.org/pdf/2509.08519
👉Project https://phantom-video.github.io/HuMo/
👉Repo https://github.com/Phantom-video/HuMo
👉Tsinghua & #ByteDance unveil HuMo: a unified, human-centric video generation framework designed to produce HQ fine-grained, and controllable human videos from multimodal inputs. It supports strong text prompt following, consistent subject preservation, synchronized audio-driven motion. Repo released under Apache2.0💙
👉Review https://t.ly/3S8Yb
👉Paper https://arxiv.org/pdf/2509.08519
👉Project https://phantom-video.github.io/HuMo/
👉Repo https://github.com/Phantom-video/HuMo
This media is not supported in your browser
VIEW IN TELEGRAM
🔥Depth Anything 3 is out🔥
👉#ByteDance unveils Depth Anything 3 (DA3), a model that predicts spatially consistent geometry from arbitrary visual inputs, with or without known camera poses. Repo under Apache 2.0💙
👉Review https://t.ly/AOPu7
👉Paper arxiv.org/pdf/2511.10647
👉Project https://lnkd.in/dnByyn2z
👉Repo https://lnkd.in/daCVz_4a
👉Demo https://lnkd.in/dKUZiJt
👉#ByteDance unveils Depth Anything 3 (DA3), a model that predicts spatially consistent geometry from arbitrary visual inputs, with or without known camera poses. Repo under Apache 2.0💙
👉Review https://t.ly/AOPu7
👉Paper arxiv.org/pdf/2511.10647
👉Project https://lnkd.in/dnByyn2z
👉Repo https://lnkd.in/daCVz_4a
👉Demo https://lnkd.in/dKUZiJt
ByteDance выкатила Bernini — новую систему для генерации и редактирования видео, где MLLM-модель сначала «планирует» сцену на семантическом уровне, а уже потом диффузионная модель рендерит итоговый ролик. По сути, ИИ сначала думает, потом рисует. 🤯
Проект построен на базе WAN 2.2 и дополнительных чекпоинтах от команды ByteDance, а также моделей, указанных авторами в репозитории. Поддерживаются генерация видео, редактирование, работа с референсами и вставка объектов в готовые ролики.
Потестить можно тут:
bernini-ai.github.io
#AI #VideoGeneration #ByteDance #WAN22 #pepewtfa
• Sent via TeleFeed
Проект построен на базе WAN 2.2 и дополнительных чекпоинтах от команды ByteDance, а также моделей, указанных авторами в репозитории. Поддерживаются генерация видео, редактирование, работа с референсами и вставка объектов в готовые ролики.
Потестить можно тут:
bernini-ai.github.io
#AI #VideoGeneration #ByteDance #WAN22 #pepewtfa
• Sent via TeleFeed