8/3/2023 0 Comments Pytorch yolov3 finetune![]() ![]() RuntimeError: Caught RuntimeError in DataLoader worker process 0. My environment is python3.5.2, torch1.0.0, GPU K80, linuxĭetecting objects: 0%| | 0/619 Traceback (most recent call last):įile "E:/PyTorch-YOLOv3-master/test.py", line 98, inįile "E:/PyTorch-YOLOv3-master/test.py", line 36, in evaluateįor batch_i, (_, imgs, targets) in enumerate(tqdm.tqdm(dataloader, desc="Detecting objects")):įile "D:\Anaconda\envs\PyTorch-YOLOv3-master\lib\site-packages\tqdm\std.py", line 1108, in iterįile "D:\Anaconda\envs\PyTorch-YOLOv3-master\lib\site-packages\torch\utils\data\dataloader.py", line 345, in nextįile "D:\Anaconda\envs\PyTorch-YOLOv3-master\lib\site-packages\torch\utils\data\dataloader.py", line 856, in _next_dataįile "D:\Anaconda\envs\PyTorch-YOLOv3-master\lib\site-packages\torch\utils\data\dataloader.py", line 881, in _process_dataįile "D:\Anaconda\envs\PyTorch-YOLOv3-master\lib\site-packages\torch_utils.py", line 394, in reraise RuntimeError: cannot perform reduction function max on tensor with no elements because the operation does not have an identity Return torch._argmax(input, dim, keepdim) Loss_cls = (1 / nB) * self.ce_loss(pred_cls, torch.argmax(tcls, dim=1))įile "/usr/local/lib/python3.5/dist-packages/torch/functional.py", line 533, in argmax Pydev_imports.execfile(file, globals, locals) # execute the scriptįile "/root/.pycharm_helpers/pydev/_pydev_imps/_pydev_execfile.py", line 18, in execfileĮxec(compile(contents "\n", file, 'exec'), glob, loc)įile "/workspace/YOLOv3_pytorch/train.py", line 85, inįile "/usr/local/lib/python3.5/dist-packages/torch/nn/modules/module.py", line 489, in callįile "/workspace/YOLOv3_pytorch/models.py", line 259, in forwardįile "/workspace/YOLOv3_pytorch/models.py", line 202, in forward ![]() Globals = n(setup, None, None, is_module)įile "/root/.pycharm_helpers/pydev/pydevd.py", line 1015, in run Hi! I have some problems about this 'loss = model(imgs, targets)' in train.pyįile "/root/.pycharm_helpers/pydev/pydevd.py", line 1578, in | Metrics | YOLO Layer 0 | YOLO Layer 1 | YOLO Layer 2 | The epoch 0/50, the training is continued after more than 100 epochs: This means that the network is correct, but it doesn't converge. I convert the pth model to onnx and plot the network which is the same as that plotted from cfg file. I trained more than 100 epochs, the AP is around 28%, and the loss is around 5. the same as those in the code on github). I used COCO to train the code on Tesla V100 on Ubuntu.Īll the parameters are not changed(i.e. It achieves 57.9 AP50 in 51 ms on a Titan X, compared to 57.5 AP50 in 198 ms by RetinaNet, similar performance but 3.8× faster. 5 IOU mAP detection metric YOLOv3 is quite good. At 320 × 320 YOLOv3 runs in 22 ms at 28.2 mAP, as accurate as SSD but three times faster. It’s a little bigger than last time but more accurate. We also trained this new network that’s pretty swell. We present some updates to YOLO! We made a bunch of little design changes to make it better. Credit YOLOv3: An Incremental Improvement detect_image( model, img)įor more advanced usage look at the method's doc strings. ![]() # Runs the YOLO model on the image boxes = detect. # Load the image as an numpy array img = cv2. Import cv2 from pytorchyolo import detect, models # Load the YOLO model model = models. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |