MobileNetV3-SSD-Compact-Version
MobileNetV3-SSD-Compact-Version copied to clipboard
shaoshengsong
MobileNetV3 SSD的简洁版本
我有使用 mbv3_large.old.pth.tar 當作預訓練模型，對VOC2007進行訓練(5000epochs)。 但是使用eval.py的結果 : Mean Average Precision (mAP): 0.005 請問有可能是什麼問題?
During the train, I discovered that my own dataset is not suitable for this code. Could you tell me what the format is and how should I review the n_classes...
Could you please upload pre-trained weights into google drive for people who are not from china? (very interested in how it will perform with comparison to the new yolo v5,...
我输入的命令是python eval.py，错误如下，请问这是什么原因呢？ File "/home/chenjunsong/MobileNetV3-SSD-Compact-Version-master/model.py", line 461, in detect_objects suppress = torch.max(suppress, overlap[box] > max_overlap) RuntimeError: _th_max not supported on CUDAType for Bool
with open(os.path.join(data_folder, self.split + '_objects.json'), 'r') as j: FileNotFoundError: [Errno 2] No such file or directory: './TRAIN_objects.json' 请问在用自己数据集训练时，需要提前划分出train，val,test的json文件是吗？ 还有一些测试问题，不知道您能不能提供邮箱或者qq联系您。 我的q:212587194 谢谢！
作者您好，请问一下在VOC0712训练一个epoch大概需要多长时间，我用2*2080Ti去训练，但是100step就需要45s, 请问我在不影响整体训练的情况下应该要如何去调整呢？
楼主你好，可以采用state_dict的方式保存你已经训练好的模型吗？模型整个保存加载时会提醒AttributeError: Can't get attribute 'SSD300' on
