TypeError: zip argument #1 must support iteration

[xjtuzhjm]

我在训练结束的时候，进入 eval() 函数进行测试的时候, 出现了以下问题

return type(out)(map(gather_map, zip(*outputs)))
TypeError: zip argument #1 must support iteration

下面是训练过程的详细信息：

`INFO:root:Munch({'batch_size': 32, 'workers': 0, 'nepoch': 100, 'model_name': 'vrcnet', 'load_model': None, 'start_epoch': 0, 'num_points': 2048, 'work_dir': 'log/', 'flag': 'debug', 'loss': 'cd', 'manual_seed': None, 'use_mean_feature': False, 'step_interval_to_print': 500, 'epoch_interval_to_save': 1, 'epoch_interval_to_val': 1, 'varying_constant': '0.01, 0.1, 0.5, 1', 'varying_constant_epochs': '5, 15, 30', 'lr': 0.0001, 'lr_decay': True, 'lr_decay_interval': 40, 'lr_decay_rate': 0.7, 'lr_step_decay_epochs': None, 'lr_step_decay_rates': None, 'lr_clip': 1e-06, 'optimizer': 'Adam', 'weight_decay': 0, 'betas': '0.9, 0.999', 'layers': '1, 1, 1, 1', 'distribution_loss': 'KLD', 'knn_list': '16', 'pk': 10, 'local_folding': True, 'points_label': True, 'num_coarse_raw': 1024, 'num_fps': 2048, 'num_coarse': 2048, 'save_vis': False, 'eval_emd': False})
(62400, 2048, 3)
(2400, 2048, 3) (62400,)
(41600, 2048, 3)
(1600, 2048, 3) (41600,)
INFO:root:Length of train dataset:62400
INFO:root:Length of test dataset:41600
INFO:root:Random Seed: 785
Jitting Chamfer 3D
Loaded JIT 3D CUDA chamfer distance
Loaded JIT 3D CUDA emd
INFO:root:vrcnet_cd_debug_2021-08-08T14:50:26 train [0: 0/1950]  loss_type: cd, fine_loss: 0.183416 total_loss: 4.883119 lr: 0.000100 alpha: 0.01
INFO:root:vrcnet_cd_debug_2021-08-08T14:50:26 train [0: 500/1950]  loss_type: cd, fine_loss: 0.041089 total_loss: 0.644301 lr: 0.000100 alpha: 0.01
INFO:root:vrcnet_cd_debug_2021-08-08T14:50:26 train [0: 1000/1950]  loss_type: cd, fine_loss: 0.039373 total_loss: 0.594741 lr: 0.000100 alpha: 0.01
INFO:root:vrcnet_cd_debug_2021-08-08T14:50:26 train [0: 1500/1950]  loss_type: cd, fine_loss: 0.034346 total_loss: 0.527504 lr: 0.000100 alpha: 0.01
INFO:root:Saving net...
INFO:root:Testing...
Traceback (most recent call last):
  File "/home/zhjp/project/MVP_Benchmark/completion/train.py", line 214, in <module>
    train()
  File "/home/zhjp/project/MVP_Benchmark/completion/train.py", line 153, in train
    val(net, epoch, val_loss_meters, dataloader_test, best_epoch_losses)
  File "/home/zhjp/project/MVP_Benchmark/completion/train.py", line 171, in val
    result_dict = net(inputs, gt, prefix="val")
  File "/home/zhjp/miniconda3/envs/mvp/lib/python3.7/site-packages/torch/nn/modules/module.py", line 550, in __call__
    result = self.forward(*input, **kwargs)
  File "/home/zhjp/miniconda3/envs/mvp/lib/python3.7/site-packages/torch/nn/parallel/data_parallel.py", line 156, in forward
    return self.gather(outputs, self.output_device)
  File "/home/zhjp/miniconda3/envs/mvp/lib/python3.7/site-packages/torch/nn/parallel/data_parallel.py", line 168, in gather
    return gather(outputs, output_device, dim=self.dim)
  File "/home/zhjp/miniconda3/envs/mvp/lib/python3.7/site-packages/torch/nn/parallel/scatter_gather.py", line 68, in gather
    res = gather_map(outputs)
  File "/home/zhjp/miniconda3/envs/mvp/lib/python3.7/site-packages/torch/nn/parallel/scatter_gather.py", line 62, in gather_map
    for k in out))
  File "/home/zhjp/miniconda3/envs/mvp/lib/python3.7/site-packages/torch/nn/parallel/scatter_gather.py", line 62, in <genexpr>
    for k in out))
  File "/home/zhjp/miniconda3/envs/mvp/lib/python3.7/site-packages/torch/nn/parallel/scatter_gather.py", line 63, in gather_map
    return type(out)(map(gather_map, zip(*outputs)))
TypeError: zip argument #1 must support iteration

Process finished with exit code 1

我用四块显卡训练的，我看网上说这个问题是多显卡学习的时候的问题，但是找了好久没找到正确的解决方法。

 inputs = inputs.float().cuda()
            gt = gt.float().cuda()
            inputs = inputs.transpose(2, 1).contiguous()
            result_dict = net(inputs, gt, prefix="val")   # 就这个地方有问题
            for k, v in val_loss_meters.items():
                v.update(result_dict[k].mean().item(), curr_batch_size)

[Yuchen-Tao]

Hello, is this problem solved? I am facing the same issue when I try to train on four GPUs.

[zhujunli1993]

Me too. I also encountered the same error when I used 2 gpus during training and validating the model. Did anyone know how to fix it?

[184688164]

Perhaps you can try to change "result_dict = net(inputs, gt, prefix="val")" in train.py file to "result_dict = net.module.forward(inputs, gt, prefix="val")"