(dataset, parts=3, part_size=None)
| 90 | |
| 91 | |
| 92 | def split_dataset(dataset, parts=3, part_size=None): |
| 93 | length = len(dataset) |
| 94 | each_length = length//parts |
| 95 | # if we specify a number, we use the number to split data |
| 96 | if part_size != None and part_size < each_length: |
| 97 | each_length = part_size |
| 98 | torch.manual_seed(0) |
| 99 | train_, inference_, test_, _ = torch.utils.data.random_split(dataset, |
| 100 | [each_length, each_length, each_length, len(dataset)-(each_length*parts)]) |
| 101 | return train_, inference_, test_ |
| 102 | |
| 103 | |
| 104 | def prepare_inference_dataset(dataset): |
nothing calls this directly
no outgoing calls
no test coverage detected