Difference between revisions of "TPU v4"

From wikieduonline

Jump to navigation Jump to search

Latest revision as of 14:09, 8 April 2023

https://cloud.google.com/tpu/docs/system-architecture-tpu-vm#tpu_v4

Peak compute per Pod: 1.1 exaflops (bf16 or int8)

JAX, PyTorch, and TensorFlow

TPU, Cloud TPU, Nvidia Tensor Cores
FLOP, Exaflop, Teraflop, TFLOP, GFLOP

Retrieved from "https://www.wikieduonline.com/index.php?title=TPU_v4&oldid=241197"

Advertising: