Anh Tran

Principal Researcher at Qualcomm AI Research

[Anh Tran @anhttran_usc]

Our SwiftBrush v2 model is released on Github:
https://t.co/dnl7pWrjG7

[Anh Tran @anhttran_usc]

I officially join Qualcomm AI Research

[TechCrunch @TechCrunch]

Qualcomm acquires generative AI division of Vietnamese startup VinAI https://t.co/cwpxunufEX

[Khoi Nguyen @ngducminhkhoi]

🚨 Exciting news! 🚨

Our 2nd workshop on “Harnessing Generative Models for Synthetic Visual Datasets, SyntaGen” is accepted to #CVPR2025! 🎉 @CVPR.

Ready for cutting-edge discussions on generative models and synthetic data.

More info: https://t.co/qXZBj4h2qi (to be updated)

[Anh Tran @anhttran_usc]

Thank you @AK for sharing our work!

🖌��� #SwiftBrushV2 just painted its way into #ECCV2024!
📷 FID 8.14 on MS-COCO
📷 Instant art (0.09s on NVIDIA A100)
📷 and its arxiv released today: https://t.co/47AHUCffDG
𝐂𝐨𝐦𝐢𝐧𝐠 𝐒𝐨𝐨𝐧: evaluation code and model
#OneStep #T2I #VinAI

[AK @_akhaliq]

SwiftBrush v2

Make Your One-step Diffusion Model Better Than Its Teacher

discuss: https://t.co/8HkLuxmX8e

In this paper, we aim to enhance the performance of SwiftBrush, a prominent one-step text-to-image diffusion model, to be competitive with its multi-step Stable Diffusion counterpart. Initially, we explore the quality-diversity trade-off between SwiftBrush and SD Turbo: the former excels in image diversity, while the latter excels in image quality. This observation motivates our proposed modifications in the training methodology, including better weight initialization and efficient LoRA training. Moreover, our introduction of a novel clamped CLIP loss enhances image-text alignment and results in improved image quality. Remarkably, by combining the weights of models trained with efficient LoRA and full training, we achieve a new state-of-the-art one-step diffusion model, achieving an FID of 8.14 and surpassing all GAN-based and multi-step Stable Diffusion models.

[AK @_akhaliq]

SwiftBrush v2

Make Your One-step Diffusion Model Better Than Its Teacher

discuss: https://t.co/8HkLuxmX8e

In this paper, we aim to enhance the performance of SwiftBrush, a prominent one-step text-to-image diffusion model, to be competitive with its multi-step Stable Diffusion counterpart. Initially, we explore the quality-diversity trade-off between SwiftBrush and SD Turbo: the former excels in image diversity, while the latter excels in image quality. This observation motivates our proposed modifications in the training methodology, including better weight initialization and efficient LoRA training. Moreover, our introduction of a novel clamped CLIP loss enhances image-text alignment and results in improved image quality. Remarkably, by combining the weights of models trained with efficient LoRA and full training, we achieve a new state-of-the-art one-step diffusion model, achieving an FID of 8.14 and surpassing all GAN-based and multi-step Stable Diffusion models.

[Anh Tran @anhttran_usc]

AI-based All-Weather Surveillance System Workshop (AWSS) is calling for paper submission.

Site: https://t.co/lGKGYCthW3
Paper call: https://t.co/1OGYGfNnAE
Deadline: Aug 30

#ACCV2024 #Workshop #AWSS

[Anh Tran @anhttran_usc]

Trustworthy Machine Learning towards Advanced Vision Systems workshop (TMLAVS) is calling for paper submissions.

Site: https://t.co/oN6Mf79Ge3
Paper call: https://t.co/3nR2AZPDa0
Deadline: Aug 28

#ACCV2024 #Workshop #TMLAVS

[Anh Tran @anhttran_usc]

Large Vision – Language Model Learning and Applications workshop (LAVA) is calling for papers and challenge.

Site: https://t.co/qUblchvYt0
Paper call: https://t.co/qvItfOUMak
Challenge: https://t.co/c0m7qgTGde
Deadline (both): Sep 30

#ACCV2024 #Workshop #LAVA

[Anh Tran @anhttran_usc]

Multispectral Imaging for Robotics and Automation workshop (MIRA) is calling for papers.

Site: https://t.co/HYOZ3X3uil
Paper call: https://t.co/HYOZ3X3uil
Deadline: Sep 14

#ACCV2024 #Workshop #MiRA

[Anh Tran @anhttran_usc]

Machine Learning and Computing for Visual Semantic Analysis workshop (MLCSA) is calling for paper submissions.

Site: https://t.co/Ep4NZYwuSs
Paper call: https://t.co/6teXCLXrlS
Deadline: Sep 28

#ACCV2024 #Workshop #MLCSA

[Anh Tran @anhttran_usc]

Computer Vision for Developing Countries workshop (CV4DC) is calling for paper submissions.

Site: https://t.co/TVDVoTvdIb
Paper call: https://t.co/e4n2OpfcaH
Deadline: Sep 24

#ACCV2024 #Workshop #CV4DC

[Anh Tran @anhttran_usc]

Generative AI for Synthetic Medical Data workshop (GAISynMeD) is calling for papers.

Site: https://t.co/KHygLdjvUD
Paper deadline: Sep 20

#ACCV2024 #Workshop #GÁIynMeD

[Anh Tran @anhttran_usc]

Rich Media with Generative AI workshop (RichMediaGAI) is calling for paper and challenge.

Site: https://t.co/AjsTtVyqtK
Paper call: https://t.co/jIF1Ujj1u8
Challenge: https://t.co/Z9wnqh35hh
Deadline (both): Sep 23

#ACCV2024 #Workshop #RichMediaGAI

[Anh Tran @anhttran_usc]

Object Instance Detection workshop (InsDet) is calling for a challenge.

Site: https://t.co/mTAppY4vDL
Challenge: https://t.co/sVbDormz20
Test Phase: 10/20-11/30

#ACCV2024 #Workshop #InsDet

[Anh Tran @anhttran_usc]

Women in Computer Vision workshop (WiCV) is calling for paper submissions.

Site: https://t.co/J0LvpsNUVB
Paper call (draft): https://t.co/BNs0QAPmNq
Deadline: Sep 11

#ACCV2024 #Workshop #WiCV

[Anh Tran @anhttran_usc]

Robust, Trustworthy and Cost-Optimized Learning Across Multiple Modalities: Theory, Algorithms, and Applications (LAMM) is calling for paper submission.

Site and paper call: https://t.co/kPmDtDUVm4
Deadline: Sep 11

#ACCV2024 #Workshop #LAMM

[Anh Tran @anhttran_usc]

@ACCVConf
Robust, Trustworthy and Cost-Optimized Learning Across Multiple Modalities: Theory, Algorithms, and Applications (LAMM) is calling for paper submission.

Site and paper call: https://t.co/kPmDtDUVm4
Deadline: Sep 11

#ACCV2024 #Workshop #LAMM

[Anh Tran @anhttran_usc]

@ACCVConf
Women in Computer Vision workshop (WiCV) is calling for paper submissions.

Site: https://t.co/ErypPzyNig
Paper call (draft): https://t.co/BNs0QAPmNq
Deadline: Sep 11

#ACCV2024 #Workshop #WiCV

[Anh Tran @anhttran_usc]

@ACCVConf
Object Instance Detection workshop (InsDet) is calling for a challenge.

Site: https://t.co/hBxDguQcuS
Challenge: https://t.co/sVbDormz20
Test Phase: 10/20-11/30

#ACCV2024 #Workshop #InsDet

[Anh Tran @anhttran_usc]

@ACCVConf
Rich Media with Generative AI workshop (RichMediaGAI) is calling for paper and challenge.

Site: https://t.co/DuLpGTxGoF
Paper call: https://t.co/jIF1Ujj1u8
Challenge: https://t.co/Z9wnqh35hh
Deadline (both): Sep 23

#ACCV2024 #Workshop #RichMediaGAI