onnx input name: images, output name: output
ori_img shape: 3, 640, 640
Traceback (most recent call last):
File "detect_onnx.py", line 150, in
detect(cfg)
File "detect_onnx.py", line 95, in detect
[(src.shape[1], src.shape[0])], Ks, encoder_decoder)
File "/home/a/Documents/lzp666/Yolo3D-master/postprocess/postprocess.py", line 192, in decode_pred_logits_onnx
det = preds[indices][:, 1:]
IndexError: The shape of the mask [1, 20, 20, 22] at index 1 does not match the shape of the indexed tensor [1, 3, 20, 20, 22] at index 1
命令: python detect_onnx.py --onnx ./weights/model3d_5m_best_transconv_11_25.onnx --img-size 640 640 --save-txt 终端: Namespace(agnostic_nms=False, classes=None, conf_thres=0.4, data='./datasets/configs/kitti.yaml', device='cpu', fourcc='mp4v', img_size=[640, 640], iou_thres=0.5, is_mosaic=False, is_rect=False, onnx='./weights/model3d_5m_best_transconv_11_25.onnx', output='inference/output', save_txt=True, weights='weights/model3d_5m_best_transconv_11_25.pt', write_video=False) Using CPU
onnx input name: images, output name: output ori_img shape: 3, 640, 640 Traceback (most recent call last): File "detect_onnx.py", line 150, in
detect(cfg)
File "detect_onnx.py", line 95, in detect
[(src.shape[1], src.shape[0])], Ks, encoder_decoder)
File "/home/a/Documents/lzp666/Yolo3D-master/postprocess/postprocess.py", line 192, in decode_pred_logits_onnx
det = preds[indices][:, 1:]
IndexError: The shape of the mask [1, 20, 20, 22] at index 1 does not match the shape of the indexed tensor [1, 3, 20, 20, 22] at index 1