您好，使用我自己的coco数据集训练时报错，这个问题怎么解决呢

[dqq813]

Random Seed: 123 loading annotations into memory... Done (t=0.16s) creating index... index created! loading annotations into memory... Done (t=0.75s) creating index... index created! loading annotations into memory... Done (t=0.21s) creating index... index created! train: Epoch: 0 eta: 0:04:35 epoch: 0 step: 100 mask_loss: 1.1265 ct_loss: 9.9765 wh_loss: 0.4440 py_loss: 2.4523 shape_loss: 0.1815 loss: 14.1808 data: 0.0084 batch: 0.5586 lr: 0.000100 max_mem: 3470 eta: 0:03:43 epoch: 0 step: 200 mask_loss: 0.7927 ct_loss: 3.3213 wh_loss: 0.3655 py_loss: 2.0679 shape_loss: 0.0886 loss: 6.6360 data: 0.0091 batch: 0.5620 lr: 0.000100 max_mem: 3983 eta: 0:02:48 epoch: 0 step: 300 mask_loss: 0.6213 ct_loss: 1.8973 wh_loss: 0.3333 py_loss: 1.8896 shape_loss: 0.0981 loss: 4.8396 data: 0.0089 batch: 0.5571 lr: 0.000100 max_mem: 3983 eta: 0:01:52 epoch: 0 step: 400 mask_loss: 0.5399 ct_loss: 1.5056 wh_loss: 0.3269 py_loss: 1.8097 shape_loss: 0.0957 loss: 4.2778 data: 0.0095 batch: 0.5528 lr: 0.000100 max_mem: 3983 eta: 0:00:57 epoch: 0 step: 500 mask_loss: 0.5813 ct_loss: 1.4105 wh_loss: 0.3498 py_loss: 1.9747 shape_loss: 0.0853 loss: 4.4016 data: 0.0090 batch: 0.5660 lr: 0.000100 max_mem: 3983 eta: 0:00:01 epoch: 0 step: 600 mask_loss: 0.5160 ct_loss: 1.2015 wh_loss: 0.3157 py_loss: 1.6819 shape_loss: 0.0782 loss: 3.7933 data: 0.0087 batch: 0.5563 lr: 0.000100 max_mem: 3983 eta: 0:00:00 epoch: 0 step: 601 mask_loss: 0.5110 ct_loss: 1.2021 wh_loss: 0.3062 py_loss: 1.6261 shape_loss: 0.0758 loss: 3.7213 data: 0.0089 batch: 0.5551 lr: 0.000100 max_mem: 3983 train: Epoch: 1 eta: 0:04:41 epoch: 1 step: 702 mask_loss: 0.5021 ct_loss: 1.1397 wh_loss: 0.3219 py_loss: 1.7633 shape_loss: 0.0737 loss: 3.8006 data: 0.0090 batch: 0.5616 lr: 0.000100 max_mem: 3983 eta: 0:03:45 epoch: 1 step: 802 mask_loss: 0.4449 ct_loss: 0.9661 wh_loss: 0.2698 py_loss: 1.3796 shape_loss: 0.0610 loss: 3.1214 data: 0.0095 batch: 0.5578 lr: 0.000100 max_mem: 3983 eta: 0:02:49 epoch: 1 step: 902 mask_loss: 0.4310 ct_loss: 0.9317 wh_loss: 0.2606 py_loss: 1.3552 shape_loss: 0.0574 loss: 3.0360 data: 0.0093 batch: 0.5731 lr: 0.000100 max_mem: 3983 eta: 0:01:53 epoch: 1 step: 1002 mask_loss: 0.4756 ct_loss: 0.9769 wh_loss: 0.2400 py_loss: 1.2478 shape_loss: 0.0495 loss: 2.9897 data: 0.0104 batch: 0.5537 lr: 0.000100 max_mem: 4003 eta: 0:00:57 epoch: 1 step: 1102 mask_loss: 0.4366 ct_loss: 0.9670 wh_loss: 0.2900 py_loss: 1.4717 shape_loss: 0.0617 loss: 3.2271 data: 0.0092 batch: 0.5532 lr: 0.000100 max_mem: 4003 eta: 0:00:01 epoch: 1 step: 1202 mask_loss: 0.4228 ct_loss: 0.8414 wh_loss: 0.2606 py_loss: 1.3517 shape_loss: 0.0602 loss: 2.9368 data: 0.0092 batch: 0.5544 lr: 0.000100 max_mem: 4003 eta: 0:00:00 epoch: 1 step: 1203 mask_loss: 0.4333 ct_loss: 0.8418 wh_loss: 0.2695 py_loss: 1.4012 shape_loss: 0.0611 loss: 3.0070 data: 0.0095 batch: 0.5544 lr: 0.000100 max_mem: 4003 5%|██████▎ | 17/329 [00:04<01:01, 5.09it/s] Traceback (most recent call last): File "train_net.py", line 83, in main() File "train_net.py", line 74, in main train(cfg, network) File "train_net.py", line 59, in train trainer.val(epoch, val_loader, evaluator, recorder) File "/home/dqq/project/PolySnake/lib/train/trainers/trainer.py", line 82, in val output, loss, loss_stats, image_stats = self.network(batch) File "/home/dqq/anaconda3/envs/snake/lib/python3.7/site-packages/torch/nn/modules/module.py", line 493, in call result = self.forward(*input, kwargs) File "/home/dqq/anaconda3/envs/snake/lib/python3.7/site-packages/torch/nn/parallel/data_parallel.py", line 150, in forward return self.module(*inputs[0], *kwargs[0]) File "/home/dqq/anaconda3/envs/snake/lib/python3.7/site-packages/torch/nn/modules/module.py", line 493, in call result = self.forward(input, kwargs) File "lib/train/trainers/snake.py", line 24, in forward output = self.net(batch['inp'], batch) File "/home/dqq/anaconda3/envs/snake/lib/python3.7/site-packages/torch/nn/modules/module.py", line 493, in call result = self.forward(*input, kwargs) File "lib/networks/snake/ct_snake.py", line 22, in forward output = self.raft(output, cnn_feature, batch) File "/home/dqq/anaconda3/envs/snake/lib/python3.7/site-packages/torch/nn/modules/module.py", line 493, in call result = self.forward(*input, *kwargs) File "lib/networks/snake/ICD.py", line 96, in forward net, offset = self.update_block(net, i_poly_fea) File "/home/dqq/anaconda3/envs/snake/lib/python3.7/site-packages/torch/nn/modules/module.py", line 493, in call result = self.forward(input, kwargs) File "lib/networks/snake/update.py", line 35, in forward net = self.gru(net, i_poly_fea) File "/home/dqq/anaconda3/envs/snake/lib/python3.7/site-packages/torch/nn/modules/module.py", line 493, in call result = self.forward(*input, *kwargs) File "lib/networks/snake/update.py", line 14, in forward z = torch.sigmoid(self.convz(hx)) File "/home/dqq/anaconda3/envs/snake/lib/python3.7/site-packages/torch/nn/modules/module.py", line 493, in call result = self.forward(input, **kwargs) File "/home/dqq/anaconda3/envs/snake/lib/python3.7/site-packages/torch/nn/modules/conv.py", line 196, in forward self.padding, self.dilation, self.groups) RuntimeError: Given groups=1, weight of size 64 128 3, expected input[0, 2, 2] to have 128 channels, but got 2 channels instead

[uu88s]

It could be that your data set is not formatted correctly

[LiuReboot]

请问正确的coco数据集格式是怎样的呢？能给出数据集的目录结构吗

[ZitengXue]

您好我也遇到了这个问题，请问您解决了吗？