nerdexam
Huawei

H12-921_V1.0 · Question #30

Which of the following descriptions of distributed training is wrong?

The correct answer is D. In model parallelism, each compute node will only interact with the trained sub-model and obtain. See the full explanation below for the reasoning.

Question

Which of the following descriptions of distributed training is wrong?

Options

  • AIn data parallelism, each computing node trains the complete model based on local data
  • BDistributed training can support the training of massive data and complex models
  • CData parallelism and model parallelism are common distributed training methods
  • DIn model parallelism, each compute node will only interact with the trained sub-model and obtain

How the community answered

(27 responses)
  • A
    11% (3)
  • B
    4% (1)
  • C
    4% (1)
  • D
    81% (22)

Community Discussion

No community discussion yet for this question.

Full H12-921_V1.0 Practice