Utility function to resample the loaded datalist for training, for example: If factor < 1.0, randomly pick part of the datalist and set to Dataset, useful to quickly test the program. If factor > 1.0, repeat the datalist to enhance the Dataset. Args: data: original datalist
(data: Sequence, factor: float, random_pick: bool = False, seed: int = 0)
| 1308 | |
| 1309 | |
| 1310 | def resample_datalist(data: Sequence, factor: float, random_pick: bool = False, seed: int = 0): |
| 1311 | """ |
| 1312 | Utility function to resample the loaded datalist for training, for example: |
| 1313 | If factor < 1.0, randomly pick part of the datalist and set to Dataset, useful to quickly test the program. |
| 1314 | If factor > 1.0, repeat the datalist to enhance the Dataset. |
| 1315 | |
| 1316 | Args: |
| 1317 | data: original datalist to scale. |
| 1318 | factor: scale factor for the datalist, for example, factor=4.5, repeat the datalist 4 times and plus |
| 1319 | 50% of the original datalist. |
| 1320 | random_pick: whether to randomly pick data if scale factor has decimal part. |
| 1321 | seed: random seed to randomly pick data. |
| 1322 | |
| 1323 | """ |
| 1324 | scale, repeats = math.modf(factor) |
| 1325 | ret: list = list() |
| 1326 | |
| 1327 | for _ in range(int(repeats)): |
| 1328 | ret.extend(list(deepcopy(data))) |
| 1329 | if scale > 1e-6: |
| 1330 | ret.extend(partition_dataset(data=data, ratios=[scale, 1 - scale], shuffle=random_pick, seed=seed)[0]) |
| 1331 | |
| 1332 | return ret |
| 1333 | |
| 1334 | |
| 1335 | def select_cross_validation_folds(partitions: Sequence[Iterable], folds: Sequence[int] | int) -> list: |
searching dependent graphs…