0

尝试使用自定义图像训练模型以识别车牌字符，但这是错误：

model = model.train()

loss_history = []
for epoch in range(10):
    pbar = tqdm(dataloader_train)
    for batch_image, batch_carnumber in pbar:
        batch_image =  batch_image.float()
        batch_output = model(batch_image).squeeze()
        loss = criterion(batch_output.permute(2,0,1).log(), 
                         batch_carnumber, 
                         torch.tensor([batch_output.shape[2]]*batch_output.shape[0]), 
                         torch.tensor([batch_carnumber.shape[1]]*batch_carnumber.shape[0]))

        pbar.set_description(f'epoch {epoch}, loss:{loss.item():.4f}')
        loss_history.append(loss.item())

        optimizer.zero_grad()
        loss.backward()
        optimizer.step()

torch.save(model, 'model-lastest.pth')

输出：

  0%|          | 0/4 [00:00<?, ?it/s]
---------------------------------------------------------------------------
RuntimeError                              Traceback (most recent call last)
<ipython-input-21-f564c8f6f874> in <module>()
      4 for epoch in range(10):
      5     pbar = tqdm(dataloader_train)
----> 6     for batch_image, batch_carnumber in pbar:
      7         batch_image =  batch_image.float()
      8         batch_output = model(batch_image).squeeze()

7 frames
/usr/local/lib/python3.6/dist-packages/torch/utils/data/_utils/collate.py in default_collate(batch)
     53             storage = elem.storage()._new_shared(numel)
     54             out = elem.new(storage)
---> 55         return torch.stack(batch, 0, out=out)
     56     elif elem_type.__module__ == 'numpy' and elem_type.__name__ != 'str_' \
     57             and elem_type.__name__ != 'string_':

RuntimeError: stack expects each tensor to be equal size, but got [7] at entry 0 and [6] at entry 2

这是什么意思？也许是因为我给你训练的图像？

4

0 回答 0