Train and Test question about clip-driven-universal-model HOT 3 CLOSED

The recall for spleen, liver, postcava are all nearly 100% but other metrics are nearly 0%. It seems all ct are predicted as these three organs. Did you correctly process the label as instruction in README?

from clip-driven-universal-model.

Training Issues

Hello, may I ask how many epochs you have trained and the dice loss value at the final convergence

When I trained using three datasets (liver, kidney, and tumor), 04-LITS, 05KITS, and 10-3liver, I trained for 200 epochs. When I finally converged, the dice loss value was around 0.7. However, in the test set, the dice value range for liver and kidney was around 0.96, and the dice value for tumor was around 0.65. I don't understand why the dice loss value was so high during convergence. According to conventional understanding, when training convergence, dice predicted (1-0.7=) 0.3 correctly, The test results dice should also be very low, but the test set results are normal. Could you please answer them? Thank you very much

from clip-driven-universal-model.

Hello, I would like to ask some questions about training a new dataset. Can I reposition organ 1 to another organ? For example, the original organ 1 is the spleen, but I would like to define it as the left atrium.

from clip-driven-universal-model.