Anmol Agarwal

Anmol Agarwal

About Me

I am a UG in Electrical Engineering and Computer Science at the Indian Institute of Science, Education and Research, Bhopal.

My interests include Computer Vision, GenAI, Diffusion Models, Transformers, and Graphics Programming.

Currently exploring personalization in diffusion models and autoregressive World Models.

Experience

Research Intern - Vision and AI Lab, Indian Institute of Science

December 2024 - Ongoing

  • Working as part of the Vision and AI Lab, recognized as the #1 CV research lab in India by CSRankings.org.
  • Creating a dataset for Object Relighting using BlenderProc API.
  • Conducted a literature survey of cutting-edge CVPR2025 papers on Video Diffusion models.
  • Fine-tuned a 1.3B parameter video model on a custom dataset.
  • Analyzed state-of-the-art video diffusion models like Hunyuan, Mochi-1, and CogVideoX.
  • Compiled research findings and methodologies in a comprehensive internship report.
View Internship Report

Technical Skills

Personal Projects

Relighting videos with personalization of Wan2.1

Explored personalization techniques in video relighting using the Wan2.1 model.

DreamLab - Text2Video Generation

A Latent Diffusion Transformer-based model inspired by OpenAI's Sora, built with PyTorch.

Sim3D - Realtime Cloth Simulation

A 3D physics engine built from scratch in C++ with OpenGL, featuring cloth simulation and a scene editor.

View All Projects