训练问题

[yy11223]

yaml: !export CUDA_VISIBLE_DEVICES=0 # 设置1张可用的卡

windows下请执行以下命令

!set CUDA_VISIBLE_DEVICES=0 !cd work/PaddleSeg && python pdseg/train.py \ --cfg configs/deeplabv3p_resnet50_vd_cityscapes.yaml \ --use_gpu \ --do_eval \ 运行结果： You are using Paddle compiled with TensorRT, but TensorRT dynamic library is not found. Ignore this if TensorRT is not needed. Traceback (most recent call last): File "pdseg/train.py", line 468, in main(args) File "pdseg/train.py", line 444, in main cfg.update_from_list(args.opts) File "/home/aistudio/work/PaddleSeg/pdseg/utils/collect.py", line 140, in update_from_list format(config_list)) ValueError: Command line options config format error! Please check it: ['\'] 为何用有的骨干网络不会报错，有的会报错

[haoyuying]

您好，请问您使用的paddleseg的版本是动态图版本吗，建议使用paddleseg2.0动态图版本和我们提供的配置参数,参考配置如下： https://github.com/PaddlePaddle/PaddleSeg/blob/release/v2.0.0-rc/configs/deeplabv3p/deeplabv3p_resnet50_os8_cityscapes_1024x512_80k.yml

[yy11223]

您好， 效果可视化 当保存完模型后，可以通过PaddleSeg提供的脚本对模型预测结果进行可视化，查看分割效果。

python predict.py \ --config configs/quick_start/bisenet_optic_disc_512x512_1k.yml \ --model_path output/iter_1000/model.pdparams \ --image_path data/optic_disc_seg/JPEGImages/H0003.jpg \ --save_dir output/result 其中image_path也可以是一个目录，这时候将对目录内的所有图片进行预测并保存可视化结果图。

我设置image_path为一个目录，出错： File "predict.py", line 172, in main(args) File "predict.py", line 166, in main stride=args.stride, File "/home/aistudio/work/PaddleSeg/paddleseg/core/predict.py", line 78, in predict ori_shape = im.shape[:2] AttributeError: 'NoneType' object has no attribute 'shape' 请问如果可视化为一个目录，怎么改程序呢？

[haoyuying]

您好，--image_path data/optic_disc_seg/JPEGImages 这样尝试仍然会报错吗，这个报错原因是没有输入图片导致的，im应该是个空，所以才会报错。可以打印出https://github.com/PaddlePaddle/PaddleSeg/blob/release/v2.0.0-rc/predict.py#L151的输出，看看图片是否传入正确。

[github-actions[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed in 7 days if no further activity occurs. Thank you for your contributions.