Implement a function computing $+$ per-GPU memory for a given model/parallelism conf

Easy

Infrastructure, Distributed Training & Scaling1 / 4

Chapter 231 / 4

Infrastructure, Distributed Training & Scaling

Implement a function computing $+$ per-GPU memory for a given model/parallelism conf

Easy

Implement a function computing $+$ per-GPU memory for a given model/parallelism config.

Target interface

memory_report(num_params, dp=1, tp=1, pp=1, zero_stage=0, bytes_weight=2, bytes_grad=2, bytes_optim=12)

Implement the function/class skeleton in the editor. Any correct approach is accepted.

Hints0 / 2

Reference solution

Your own code stays in the editor.

solution.pyPython

Local draft

def memory_report(num_params, dp=1, tp=1, pp=1, zero_stage=0, bytes_weight=2, bytes_grad=2, bytes_optim=12):
    raise NotImplementedError

⌘/Ctrl↵to submit

AI review

No review yet

Submit your solution for a verdict, a score out of 100 and concrete feedback. Any correct approach passes.

Chapter 231 / 4

Infrastructure, Distributed Training & Scaling

Implement a function computing $+$ per-GPU memory for a given model/parallelism conf

Easy

Implement a function computing $+$ per-GPU memory for a given model/parallelism config.

Target interface

memory_report(num_params, dp=1, tp=1, pp=1, zero_stage=0, bytes_weight=2, bytes_grad=2, bytes_optim=12)

Implement the function/class skeleton in the editor. Any correct approach is accepted.

Hints0 / 2

Reference solution

Your own code stays in the editor.