Yongchang Hao

yongchanghao

AI & ML interests

None yet

Recent Activity

Organizations

Social Post Explorers's profile picture

yongchanghao's activity

Reacted to their post with πŸ”₯ 24 days ago
view post
Post
3733
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)
posted an update 27 days ago
view post
Post
3733
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)
New activity in EleutherAI/coqa about 1 month ago
New activity in THUDM/LongBench 2 months ago
New activity in Anthropic/hh-rlhf over 1 year ago
New activity in google/bigbench almost 2 years ago

Missing tasks

#5 opened almost 2 years ago by yongchanghao