Zephyr69

Make batchsize a multiple of gpu count and see if it works.

The coordinates are earily found in the code. But the challenge here is a clearly annotated dataset, which also implies a highly capable imaging system with extremely short exposure time....

The most background images I've worked with was a 50/50 single-class dataset, which ended up with normal metrics. If you are trying a 10/90 dataset, you definitely need to make...

Is there a workaround for this problem on Jetpack 4.6.1? I'm having the exact same issue, and migrating the entire project to ubuntu 20.04 just for trt is really not...

Building tensorrt from source resulted in the same issue too.

@zerollzeng Is there some way to downgrade to Jetpack 4.5.x?

@jylink On my NX 8gb ram EMMC module with Jetpack 4.6 and Tensorrt 8.0.1.6, it works fine. But on the 16bg ram module with the same software config as yours,...

> > @jylink On my NX 8gb ram EMMC module with Jetpack 4.6 and Tensorrt 8.0.1.6, it works fine. But on the 16bg ram module with the same software config...

For defect detection, the image space will surely be wildly different from that the pre-trained model was trained on. The pre-trained weight will be useless if not detrimental here. But...

Has there been any news on this? I still fail to figure out how to plot CI using neuralprophet :(. I have no problem with this using prophet.