X / Twitter
Pushed @gradio to its limits in one weekend hack! 🚀 Built a full web app with:
•Google OAuth
•Custom landing page
•@supabase for the DB
•@stripe for payments
•@replicate for inference/training
• @FastAPI for routes https://t.co/r1YIFdNsoB
A question I often get asked is how I raised my two daughters. People who meet them are impressed and want to know what technique we used.
The answer is that there was no time where my wife or I ever punished them. And if we ever shamed them, we apologized immediately.
This is such a cool thing to do as a dad, interviewed his daughter every first day of school https://t.co/PpX4dD2jj5
🔥IC-Light is a project built with Gradio to manipulate the illumination of images!
Project uses two types of models: text-conditioned relighting model and background-conditioned model. Both types take foreground images as inputs💡
How you can use it 👇 https://t.co/eK2fKHtuHA
This 2B pretrained Sapiens-Lite looks like a great backbone for many video tasks.
code & checkpoints: https://t.co/lDvDtIzpbt https://t.co/9nFTlVWmfb
Meta present Sapiens, a family of models for four fundamental human-centric vision tasks --
- 2D Pose Estimation,
- Body-part Segmentation,
- Depth Estimation, and
- Surface Normal Prediction https://t.co/PRA4rmk3Zs
Meta presents Sapiens
Foundation for Human Vision Models
discuss: https://t.co/LH0tEgJvnX
We present Sapiens, a family of models for four fundamental human-centric vision tasks - 2D pose estimation, body-part segmentation, depth estimation, and surface normal prediction. Our models natively support 1K high-resolution inference and are extremely easy to adapt for individual tasks by simply fine-tuning models pretrained on over 300 million in-the-wild human images. We observe that, given the same computational budget, self-supervised pretraining on a curated dataset of human images significantly boosts the performance for a diverse set of human-centric tasks. The resulting models exhibit remarkable generalization to in-the-wild data, even when labeled data is scarce or entirely synthetic. Our simple model design also brings scalability - model performance across tasks improves as we scale the number of parameters from 0.3 to 2 billion. Sapiens consistently surpasses existing baselines across various human-centric benchmarks. We achieve significant improvements over the prior state-of-the-art on Humans-5K (pose) by 7.6 mAP, Humans-2K (part-seg) by 17.1 mIoU, Hi4D (depth) by 22.4% relative RMSE, and THuman2 (normal) by 53.5% relative angular error.
Never forget one of the greatest channels of all time:
https://t.co/7v7Brb2qNG https://t.co/KCVl680lhx https://t.co/wUegnfxQdD
We have implemented an image painter using DiffSynth, which support FLUX, Kolor, HunyuanDit, Stable Diffusion series.
@_akhaliq @Gradio gradio demo is here:
https://t.co/DRmgLzsQzB
#DiffSynthPainter #InteractiveArt #InnovativePainting https://t.co/eWGT4F1USO
FancyVideo: Towards Dynamic & Consistent Video Generation via Cross-frame Textual Guidance. Project: https://t.co/xDL1fVFfnU
Gradio is 1-stop solution for production-ready GenAI apps!
- Slick UIs
- Ease of building
- Scalable
- Apps as API endpoints
https://t.co/ISNtjNvSkV
Salesforce's BLIP-3 adopts our VLFeedback for DPO training 🚀 https://t.co/TRrpiLNoDh