RuntimeError：在 DataLoader 工作进程 3 中捕获 RuntimeError

Question

我在尝试在单个 GPU 上训练模型时遇到了这个问题；我检查了张量形状，它的大小调整得很好

RuntimeError: Caught RuntimeError in DataLoader worker process 3.
Original Traceback (most recent call last):
  File "C:\ProgramData\Anaconda3\envs\new\lib\site-packages\torch\utils\data\_utils\worker.py", line 202, in _worker_loop
    data = fetcher.fetch(index)
  File "C:\ProgramData\Anaconda3\envs\new\lib\site-packages\torch\utils\data\_utils\fetch.py", line 28, in fetch
    data.append(next(self.dataset_iter))
  File "c:\users\anaszafar\desktop\train\detectron2\detectron2\data\common.py", line 201, in __iter__
    yield self.dataset[idx]
  File "c:\users\anaszafar\desktop\train\detectron2\detectron2\data\common.py", line 90, in __getitem__
    data = self._map_func(self._dataset[cur_idx])
  File "c:\users\anaszafar\desktop\train\adelaidet\adet\data\fcpose_dataset_mapper.py", line 62, in __call__
    raise
RuntimeError: No active exception to reraise

---

def __call__(self, dataset_dict):
        for i in range(100):
            dataset_dict_temp = copy.deepcopy(dataset_dict)
            dataset_dict_temp = super().__call__(dataset_dict_temp)
            if len(dataset_dict_temp["instances"]) != 0:
                if self.is_train:
                    dataset_dict_temp['instances'] = self.HeatmapGenerator(dataset_dict_temp['instances'], 
                                                                        self.gt_heatmap_stride)
                return dataset_dict_temp
        raise

Answer 1

看起来您的代码抛出了异常，但没有有意义的错误消息。查看函数 __call__ 中第 62 行附近的 c:\users\anaszafar\desktop\train\adelaidet\adet\data\fcpose_dataset_mapper.py 中的代码，看看为什么你的数据集代码决定引发一个非特定错误。