2 ms·
TPU is also slow, they used pod with 64 TPUs for training BERT. You probably can achieve similar result using distributed training on multiple GPU machines.
by riku_iki 8y ago
TPU is also slow, they used pod with 64 TPUs for training BERT.
You probably can achieve similar result using distributed training on multiple GPU machines.