README
🐠
University of Alberta
university
Verified
AI & ML interests
None defined yet.
Recent Activity
View all activity
Sagaust
published
a
Space
about 1 month ago
tafseer-nayeem
authored
2
papers
3 months ago
tafseer-nayeem
authored
a
paper
6 months ago
kgmills
authored
8
papers
7 months ago
GOAt: Explaining Graph Neural Networks via Graph Output Attribution
Paper
•
2401.14578
•
Published
•
1
Qua$^2$SeDiMo: Quantifiable Quantization Sensitivity of Diffusion Models
Paper
•
2412.14628
•
Published
Building Optimal Neural Architectures using Interpretable Knowledge
Paper
•
2403.13293
•
Published
AIO-P: Expanding Neural Performance Predictors Beyond Image Classification
Paper
•
2211.17228
•
Published
GENNAPE: Towards Generalized Neural Architecture Performance Estimators
Paper
•
2211.17226
•
Published
L$^{2}$NAS: Learning to Optimize Neural Architectures via Continuous-Action Reinforcement Learning
Paper
•
2109.12425
•
Published
Applying Graph Explanation to Operator Fusion
Paper
•
2501.00636
•
Published
FP4DiT: Towards Effective Floating Point Quantization for Diffusion Transformers
Paper
•
2503.15465
•
Published
•
1
yongchanghao
authored
a
paper
about 1 year ago
yongchanghao
posted
an
update
about 1 year ago
Post
3821
We just released a paper (NeuZip) that compresses VRAM in a lossless manner to run larger models. This should be particularly useful when VRAM is insufficient during training/inference. Specifically, we look inside each floating number and find that the exponents are highly compressible (as shown in the figure below).
Read more about the work at NeuZip: Memory-Efficient Training and Inference with Dynamic Compression of Neural Networks (2410.20650)
Read more about the work at NeuZip: Memory-Efficient Training and Inference with Dynamic Compression of Neural Networks (2410.20650)
drafiei
authored
2
papers
over 1 year ago
Philip-Huo
authored
a
paper
over 1 year ago
the-abrar-fahim
authored
a
paper
over 1 year ago
ehsk
authored
2
papers
about 2 years ago