有人遇到过上面的错误吗?知道是什么引起的吗?
我只在一个特定的数据集上得到错误，并且它在第19批加载时抛出，总是19。即使数据是随机分割的
数据集由5秒音频。wav文件组成，我在获取时将其转换为频谱图。
Starting training for 1000 epoch(s)
Epoch 1/1000
26%|█████████████████████▉ | 19/72 [00:04<00:13, 3.86it/s, loss=33.745]
Traceback (most recent call last):
File "train.py", line 202, in <module>
train_and_evaluate(model, train_dl, test_dl, optimizer, cost, params, args.model_dir, args.experiment_name, args.restore_file)
File "train.py", line 129, in train_and_evaluate
train(model, optimizer, loss_fn, train_dataloader, params, epoch, writer)
File "train.py", line 52, in train
for i, (train_batch, current_rpm, next_rpm, next_pr) in enumerate(dataloader):
File "/home/ubuntu/miniconda3/envs/SFB/lib/python3.8/site-packages/torch/utils/data/dataloader.py", line 628, in __next__
data = self._next_data()
File "/home/ubuntu/miniconda3/envs/SFB/lib/python3.8/site-packages/torch/utils/data/dataloader.py", line 1333, in _next_data
return self._process_data(data)
File "/home/ubuntu/miniconda3/envs/SFB/lib/python3.8/site-packages/torch/utils/data/dataloader.py", line 1359, in _process_data
data.reraise()
File "/home/ubuntu/miniconda3/envs/SFB/lib/python3.8/site-packages/torch/_utils.py", line 543, in reraise
raise exception
RuntimeError: Caught RuntimeError in DataLoader worker process 3.
Original Traceback (most recent call last):
File "/home/ubuntu/miniconda3/envs/SFB/lib/python3.8/site-packages/torch/utils/data/_utils/worker.py", line 302, in _worker_loop
data = fetcher.fetch(index)
File "/home/ubuntu/miniconda3/envs/SFB/lib/python3.8/site-packages/torch/utils/data/_utils/fetch.py", line 61, in fetch
return self.collate_fn(data)
File "/home/ubuntu/miniconda3/envs/SFB/lib/python3.8/site-packages/torch/utils/data/_utils/collate.py", line 265, in default_collate
return collate(batch, collate_fn_map=default_collate_fn_map)
File "/home/ubuntu/miniconda3/envs/SFB/lib/python3.8/site-packages/torch/utils/data/_utils/collate.py", line 143, in collate
return [collate(samples, collate_fn_map=collate_fn_map) for samples in transposed] # Backwards compatibility.
File "/home/ubuntu/miniconda3/envs/SFB/lib/python3.8/site-packages/torch/utils/data/_utils/collate.py", line 143, in <listcomp>
return [collate(samples, collate_fn_map=collate_fn_map) for samples in transposed] # Backwards compatibility.
File "/home/ubuntu/miniconda3/envs/SFB/lib/python3.8/site-packages/torch/utils/data/_utils/collate.py", line 120, in collate
return collate_fn_map[elem_type](batch, collate_fn_map=collate_fn_map)
File "/home/ubuntu/miniconda3/envs/SFB/lib/python3.8/site-packages/torch/utils/data/_utils/collate.py", line 162, in collate_tensor_fn
out = elem.new(storage).resize_(len(batch), *list(elem.size()))
RuntimeError: Trying to resize storage that is not resizable
我已经尝试了2秒。wav录音的不同数据集，它工作得很好。
我在为嵌入式系统开发计算机视觉项目时遇到了这个错误。该项目的目标是使用神经网络比较两幅图像，该网络必须说明这两幅图像是否相似。为了减小网络的大小，我没有将整个图像馈送给网络，而是使用ORB算法获得的特征向量。
问题是我从每个图像(不同的关键点)获得一个大小不同的向量。Dataloader对象的默认collate_fn不允许这种情况发生。我有两个解决方案:
- 手动添加一种填充，使当前批处理中的所有特征向量具有相同的大小;或
- 编写一个自定义collate_fn能够处理不同大小的批处理元素
我遵循了第一个解决方案，它对我有效:我编写了一个自定义Dataset对象，将每个特征向量调整大小并填充到固定大小(在__getitem__方法中)。
也许，在你的情况下，你应该需要一个自定义的collate_fn。
我建议你阅读:
- Dataloader类的官方文档
- 博客文章，展示了一个NLP用例，其中自定义collate_fn非常有用