Mmdetection loading from own training checkpoint for inference produces garbage detections asked 4 years, 4 months ago modified 2 years, 6 months ago viewed 5k times Mmdet has a very good community, you can try to look at this link to solve your problem. 'configdict' object has no attribute 'data' on mmdetection?
How To Remotely Access Raspberry Pi With RemoteIoT On Windows A
I was meeting the same issue while running mmdetection training job on multiple gpus. Asked 4 years, 2 months ago modified 2 years, 9 months ago viewed 2k times It will not affect the training.
Use cudatoolkit==10.0 not 10.2 so that pytorch 1.4 is used instead of pytorch 1.5.
I tried to train my model with mmdetection, however, error like keyerror: To gain full voting privileges, Shortly, you should change your classes = ('pantograph',) as the following codes. I've added my dataset to.
Tried to train the model but got the mmcv error no module named 'mmcv._ext' mmcv library is already installed and imported mmcv version = 1.4.0 cuda version = 10.0. However if using mmdetection v2.0.0 it seems that pytorch 1.5 with cudatoolkit==10.2. Asked 2 years, 10 months ago modified 1 year, 5 months ago viewed 5k times How to save the model weights after running train_detector in mmdetection?
To use 4 gpu devices (onto the same host) i've run my training script via.
It is normal that the model and loaded state dict do not match exactly, because the fully connected layers in the pretrained models are unused. 'mask_detectiondataset is not in the dataset registry' keep showing.
