`vqa_benchmarking_backend.metrics.model_info`¶

Module Contents¶

Functions¶

model_info(net: torch.nn.Module, only_trainable: bool = True)

Returns:

vqa_benchmarking_backend.metrics.model_info.model_info(net: torch.nn.Module, only_trainable: bool = True)¶: Returns: int: the total number of parameters used by net (only counting shared parameters once);

if only_trainable is True, then only counts parameters with requires_grad = True