JOHNSON LIN's repositories
sam_inpaint
SAM + Inpainting
sparsecl.github.io
demo page for the preprint: "SPARSECL: Sparse Contrastive Learning for Contradiction Retrieval"
AMT
Official code for "AMT: All-Pairs Multi-Field Transforms for Efficient Frame Interpolation" (CVPR2023)
contradiction-retrieval
NeurIPS 2024 Submission: "SPARSECL: Sparse Contrastive Learning for Contradiction Retrieval"
flash-attention
Fast and memory-efficient exact attention
Improving-T0
Official code for paper "Not All Tasks are Born Equal: Uniderstanding Zero-Shot Generalization"
Megatron-LM
Ongoing research training transformer models at scale
Cosmos-Tokenizer
A suite of image and video neural tokenizers
diffusion_policy
[RSS 2023] Diffusion Policy Visuomotor Policy Learning via Action Diffusion
vidai-johnson
Video generation using AI
videophy-benchmark
Video Generation, Physical Commonsense, Semantic Adherence, VideoCon-Physics
videophy.github.io
Demo Page for Preprint: "VideoPhy: Evaluating Physical Commonsense In Video Generation"