(self, *args, **kwargs)
| 70 | """ |
| 71 | |
| 72 | def __init__(self, *args, **kwargs): |
| 73 | super().__init__(*args, **kwargs) |
| 74 | self.__initialized = False |
| 75 | shuffle = False |
| 76 | batch_sampler = None |
| 77 | if len(args) > 5: |
| 78 | shuffle = args[2] |
| 79 | sampler = args[3] |
| 80 | batch_sampler = args[4] |
| 81 | elif len(args) > 4: |
| 82 | shuffle = args[2] |
| 83 | sampler = args[3] |
| 84 | if "batch_sampler" in kwargs: |
| 85 | batch_sampler = kwargs["batch_sampler"] |
| 86 | elif len(args) > 3: |
| 87 | shuffle = args[2] |
| 88 | if "sampler" in kwargs: |
| 89 | sampler = kwargs["sampler"] |
| 90 | if "batch_sampler" in kwargs: |
| 91 | batch_sampler = kwargs["batch_sampler"] |
| 92 | else: |
| 93 | if "shuffle" in kwargs: |
| 94 | shuffle = kwargs["shuffle"] |
| 95 | if "sampler" in kwargs: |
| 96 | sampler = kwargs["sampler"] |
| 97 | if "batch_sampler" in kwargs: |
| 98 | batch_sampler = kwargs["batch_sampler"] |
| 99 | |
| 100 | # Use custom BatchSampler |
| 101 | if batch_sampler is None: |
| 102 | if sampler is None: |
| 103 | if shuffle: |
| 104 | sampler = torch.utils.data.sampler.RandomSampler(self.dataset) |
| 105 | # sampler = torch.utils.data.DistributedSampler(self.dataset) |
| 106 | else: |
| 107 | sampler = torch.utils.data.sampler.SequentialSampler(self.dataset) |
| 108 | batch_sampler = YoloBatchSampler( |
| 109 | sampler, |
| 110 | self.batch_size, |
| 111 | self.drop_last, |
| 112 | input_dimension=self.dataset.input_dim, |
| 113 | ) |
| 114 | # batch_sampler = IterationBasedBatchSampler(batch_sampler, num_iterations = |
| 115 | |
| 116 | self.batch_sampler = batch_sampler |
| 117 | |
| 118 | self.__initialized = True |
| 119 | |
| 120 | def close_mosaic(self): |
| 121 | self.batch_sampler.mosaic = False |
nothing calls this directly
no test coverage detected