OWOD任务训练问题

[wl654655902]

run_owod.sh第1，2，3步没有问题，自定义数据集task1有25类，task2有5类，第4步训练时task1时没有问题，训练task2时出现cls_pred.permute(0, 2, 3, 1).reshape(num_imgs, -1, self.num_classes)RuntimeError: shape '[16, -1, 32]' is invalid for input of size 44800
具体如下：
Traceback (most recent call last):
File "/home/wl/code/opensource/open_detect/YOLO-UniOW/./tools/train_owod.py", line 188, in
main()
File "/home/wl/code/opensource/open_detect/YOLO-UniOW/./tools/train_owod.py", line 183, in main
model = runner.train_loop.run() # type: ignore
File "/root/anaconda3/envs/yolouniow/lib/python3.9/site-packages/mmengine/runner/loops.py", line 96, in run
self.run_epoch()
File "/root/anaconda3/envs/yolouniow/lib/python3.9/site-packages/mmengine/runner/loops.py", line 112, in run_epoch
self.run_iter(idx, data_batch)
File "/root/anaconda3/envs/yolouniow/lib/python3.9/site-packages/mmengine/runner/loops.py", line 128, in run_iter
outputs = self.runner.model.train_step(
File "/root/anaconda3/envs/yolouniow/lib/python3.9/site-packages/mmengine/model/wrappers/distributed.py", line 121, in train_step
losses = self._run_forward(data, mode='loss')
File "/root/anaconda3/envs/yolouniow/lib/python3.9/site-packages/mmengine/model/wrappers/distributed.py", line 161, in _run_forward
results = self(**data, mode=mode)
File "/root/anaconda3/envs/yolouniow/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "/root/anaconda3/envs/yolouniow/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl
return forward_call(*args, **kwargs)
File "/root/anaconda3/envs/yolouniow/lib/python3.9/site-packages/torch/nn/parallel/distributed.py", line 1519, in forward
else self._run_ddp_forward(*inputs, **kwargs)
File "/root/anaconda3/envs/yolouniow/lib/python3.9/site-packages/torch/nn/parallel/distributed.py", line 1355, in _run_ddp_forward
return self.module(*inputs, **kwargs) # type: ignore[index]
File "/root/anaconda3/envs/yolouniow/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1518, in _wrapped_call_impl
return self._call_impl(*args, **kwargs)
File "/root/anaconda3/envs/yolouniow/lib/python3.9/site-packages/torch/nn/modules/module.py", line 1527, in _call_impl
return forward_call(*args, **kwargs)
File "/root/anaconda3/envs/yolouniow/lib/python3.9/site-packages/mmdet/models/detectors/base.py", line 92, in forward
return self.loss(inputs, data_samples)
File "/home/wl/code/opensource/open_detect/YOLO-UniOW/yolo_world/models/detectors/yolo_world_owod.py", line 95, in loss
losses = self.bbox_head.loss(img_feats, txt_feats,
File "/home/wl/code/opensource/open_detect/YOLO-UniOW/yolo_world/models/dense_heads/yolov10_world_head.py", line 376, in loss
losses = self.loss_by_feat([one2many_loss_inputs, one2one_loss_inputs])
File "/home/wl/code/opensource/open_detect/YOLO-UniOW/yolo_world/models/dense_heads/yolov10_world_head.py", line 471, in loss_by_feat
losses = self.one2many_loss_by_feat(*one2many_loss_inputs)
File "/home/wl/code/opensource/open_detect/YOLO-UniOW/third_party/mmyolo/mmyolo/models/dense_heads/yolov10_head.py", line 542, in one2many_loss_by_feat
flatten_cls_preds = [
File "/home/wl/code/opensource/open_detect/YOLO-UniOW/third_party/mmyolo/mmyolo/models/dense_heads/yolov10_head.py", line 543, in
cls_pred.permute(0, 2, 3, 1).reshape(num_imgs, -1, self.num_classes)
RuntimeError: shape '[16, -1, 32]' is invalid for input of size 44800

[wl654655902]

我数据类别这样设置的：
owod_settings = {
"KOALA": {
"task_list": [0, 25, 30],
"test_image_set": "all_task_test"
}}

[wl654655902]

问题已解决