search

Media-Smart / vedadet

A single stage object detection toolbox based on PyTorch
Apache License 2.0
498 stars 128 forks source link

Can't train in colab #73

Open HiImBug opened 2 years ago

HiImBug commented 2 years ago

Hi thank you for sharing I'm trying to train in colab but a have problem, somebody can help me pls? I'm sure the folder structure as author. I have tested in colab the following versions : CUDA: 11.2 PyTorch 1.6.0 Python 3.8.5

!CUDA_VISIBLE_DEVICES="0" python tools/trainval.py configs/trainval/tinaface/tinaface_r50_fpn_bn.py

unexpected key in source state_dict: backbone.fc.weight, backbone.fc.bias

missing keys in source state_dict: neck.0.lateral_convs.0.conv.weight, neck.0.lateral_convs.0.bn.weight, neck.0.lateral_convs.0.bn.bias, neck.0.lateral_convs.0.bn.running_mean, neck.0.lateral_convs.0.bn.running_var, neck.0.lateral_convs.1.conv.weight, neck.0.lateral_convs.1.bn.weight, neck.0.lateral_convs.1.bn.bias, neck.0.lateral_convs.1.bn.running_mean, neck.0.lateral_convs.1.bn.running_var, neck.0.lateral_convs.2.conv.weight, neck.0.lateral_convs.2.bn.weight, neck.0.lateral_convs.2.bn.bias, neck.0.lateral_convs.2.bn.running_mean, neck.0.lateral_convs.2.bn.running_var, neck.0.lateral_convs.3.conv.weight, neck.0.lateral_convs.3.bn.weight, neck.0.lateral_convs.3.bn.bias, neck.0.lateral_convs.3.bn.running_mean, neck.0.lateral_convs.3.bn.running_var, neck.0.fpn_convs.0.conv.weight, neck.0.fpn_convs.0.bn.weight, neck.0.fpn_convs.0.bn.bias, neck.0.fpn_convs.0.bn.running_mean, neck.0.fpn_convs.0.bn.running_var, neck.0.fpn_convs.1.conv.weight, neck.0.fpn_convs.1.bn.weight, neck.0.fpn_convs.1.bn.bias, neck.0.fpn_convs.1.bn.running_mean, neck.0.fpn_convs.1.bn.running_var, neck.0.fpn_convs.2.conv.weight, neck.0.fpn_convs.2.bn.weight, neck.0.fpn_convs.2.bn.bias, neck.0.fpn_convs.2.bn.running_mean, neck.0.fpn_convs.2.bn.running_var, neck.0.fpn_convs.3.conv.weight, neck.0.fpn_convs.3.bn.weight, neck.0.fpn_convs.3.bn.bias, neck.0.fpn_convs.3.bn.running_mean, neck.0.fpn_convs.3.bn.running_var, neck.0.fpn_convs.4.conv.weight, neck.0.fpn_convs.4.bn.weight, neck.0.fpn_convs.4.bn.bias, neck.0.fpn_convs.4.bn.running_mean, neck.0.fpn_convs.4.bn.running_var, neck.0.fpn_convs.5.conv.weight, neck.0.fpn_convs.5.bn.weight, neck.0.fpn_convs.5.bn.bias, neck.0.fpn_convs.5.bn.running_mean, neck.0.fpn_convs.5.bn.running_var, neck.1.level_convs.0.0.conv.weight, neck.1.level_convs.0.0.bn.weight, neck.1.level_convs.0.0.bn.bias, neck.1.level_convs.0.0.bn.running_mean, neck.1.level_convs.0.0.bn.running_var, neck.1.level_convs.0.1.conv.weight, neck.1.level_convs.0.1.bn.weight, neck.1.level_convs.0.1.bn.bias, neck.1.level_convs.0.1.bn.running_mean, neck.1.level_convs.0.1.bn.running_var, neck.1.level_convs.0.2.conv.weight, neck.1.level_convs.0.2.bn.weight, neck.1.level_convs.0.2.bn.bias, neck.1.level_convs.0.2.bn.running_mean, neck.1.level_convs.0.2.bn.running_var, neck.1.level_convs.0.3.conv.weight, neck.1.level_convs.0.3.bn.weight, neck.1.level_convs.0.3.bn.bias, neck.1.level_convs.0.3.bn.running_mean, neck.1.level_convs.0.3.bn.running_var, neck.1.level_convs.0.4.conv.weight, neck.1.level_convs.0.4.bn.weight, neck.1.level_convs.0.4.bn.bias, neck.1.level_convs.0.4.bn.running_mean, neck.1.level_convs.0.4.bn.running_var, bbox_head.cls_convs.0.conv.weight, bbox_head.cls_convs.0.bn.weight, bbox_head.cls_convs.0.bn.bias, bbox_head.cls_convs.0.bn.running_mean, bbox_head.cls_convs.0.bn.running_var, bbox_head.cls_convs.1.conv.weight, bbox_head.cls_convs.1.bn.weight, bbox_head.cls_convs.1.bn.bias, bbox_head.cls_convs.1.bn.running_mean, bbox_head.cls_convs.1.bn.running_var, bbox_head.cls_convs.2.conv.weight, bbox_head.cls_convs.2.bn.weight, bbox_head.cls_convs.2.bn.bias, bbox_head.cls_convs.2.bn.running_mean, bbox_head.cls_convs.2.bn.running_var, bbox_head.cls_convs.3.conv.weight, bbox_head.cls_convs.3.bn.weight, bbox_head.cls_convs.3.bn.bias, bbox_head.cls_convs.3.bn.running_mean, bbox_head.cls_convs.3.bn.running_var, bbox_head.reg_convs.0.conv.weight, bbox_head.reg_convs.0.bn.weight, bbox_head.reg_convs.0.bn.bias, bbox_head.reg_convs.0.bn.running_mean, bbox_head.reg_convs.0.bn.running_var, bbox_head.reg_convs.1.conv.weight, bbox_head.reg_convs.1.bn.weight, bbox_head.reg_convs.1.bn.bias, bbox_head.reg_convs.1.bn.running_mean, bbox_head.reg_convs.1.bn.running_var, bbox_head.reg_convs.2.conv.weight, bbox_head.reg_convs.2.bn.weight, bbox_head.reg_convs.2.bn.bias, bbox_head.reg_convs.2.bn.running_mean, bbox_head.reg_convs.2.bn.running_var, bbox_head.reg_convs.3.conv.weight, bbox_head.reg_convs.3.bn.weight, bbox_head.reg_convs.3.bn.bias, bbox_head.reg_convs.3.bn.running_mean, bbox_head.reg_convs.3.bn.running_var, bbox_head.retina_cls.weight, bbox_head.retina_cls.bias, bbox_head.retina_reg.weight, bbox_head.retina_reg.bias, bbox_head.retina_iou.weight, bbox_head.retina_iou.bias

Traceback (most recent call last): File "tools/trainval.py", line 65, in main() File "tools/trainval.py", line 61, in main trainval(cfg, distributed, logger) File "/content/drive/My Drive/Colab Notebooks/vedadet/vedadet/assembler/trainval.py", line 86, in trainval looper.start(cfg.max_epochs) File "/content/drive/My Drive/Colab Notebooks/vedadet/vedacore/loopers/epoch_based_looper.py", line 29, in start self.epoch_loop(mode) File "/content/drive/My Drive/Colab Notebooks/vedadet/vedacore/loopers/epoch_based_looper.py", line 15, in epoch_loop for idx, data in enumerate(dataloader): File "/usr/local/lib/python3.7/dist-packages/torch/utils/data/dataloader.py", line 359, in iter return self._get_iterator() File "/usr/local/lib/python3.7/dist-packages/torch/utils/data/dataloader.py", line 305, in _get_iterator return _MultiProcessingDataLoaderIter(self) File "/usr/local/lib/python3.7/dist-packages/torch/utils/data/dataloader.py", line 944, in init self._reset(loader, first_iter=True) File "/usr/local/lib/python3.7/dist-packages/torch/utils/data/dataloader.py", line 975, in _reset self._try_put_index() File "/usr/local/lib/python3.7/dist-packages/torch/utils/data/dataloader.py", line 1209, in _try_put_index index = self._next_index() File "/usr/local/lib/python3.7/dist-packages/torch/utils/data/dataloader.py", line 512, in _next_index return next(self._sampler_iter) # may raise StopIteration File "/usr/local/lib/python3.7/dist-packages/torch/utils/data/sampler.py", line 226, in iter for idx in self.sampler: File "/content/drive/My Drive/Colab Notebooks/vedadet/vedadet/datasets/samplers/group_sampler.py", line 39, in iter indices = np.concatenate(indices) File "<__array_function__ internals>", line 6, in concatenate ValueError: need at least one array to concatenate

HiImBug commented 2 years ago

I test so there is one folder image and annotations image image image

HiImBug commented 2 years ago

self.group_sizes = [] ('self.flag', array([], dtype=int64)) what is self.flag = self.dataset.flag in line 91 /vedadet/vedadet/datasets/samplers/group_sampler.py i can't find any field flag in file config data tinaface_r50_fpn_bn.py or tinaface_r50_fpn_gn_dcn.py in File "/content/drive/My Drive/Colab Notebooks/vedadet/vedadet/datasets/samplers/group_sampler.py", line 41, in iter