darkz baron's picture

darkz baron

darkzbaron

·

AI & ML interests

nlp and cv and ai

Recent Activity

liked a Space about 14 hours ago

bytedance-research/UNO-FLUX

reacted to onekq's post with 🚀 about 17 hours ago

We desperately need GPU for model inference. CPU can't replace GPU. I will start with the basics. GPU is designed to serve predictable workloads with many parallel units (pixels, tensors, tokens). So a GPU allocates as much transistor budget as possible to build thousands of compute units (Cuda cores in NVidia or execution units in Apple Silicon), each capable of running a thread. But CPU is designed to handle all kinds of workloads. CPU cores are much larger (hence a lot fewer) with branch prediction and other complex things. In addition, more and more transistors are allocated to build larger cache (~50% now) to house the unpredictable, devouring the compute budget. Generalists can't beat specialists.

upvoted a paper 3 days ago

Challenging BIG-Bench Tasks and Whether Chain-of-Thought Can Solve Them

View all activity

Organizations

None yet

spaces 1

No application file

Darkzbaron Public

models

None public yet

datasets

None public yet