Open rezachu opened 12 months ago
The previous issue card concluded that the model did not train toward the loss. I have been tuning the loss function and not getting a good result. Wonder is there any recommendation for tuning the loss function setup. Many Thanks.
Hi, I don't know if you are referring to the adjustment of the loss function with a plus or minus sign or? Theoretically, our loss function doesn't need to be adjusted, you can preferably try to change the hyperparameters for model optimization.
The previous issue card concluded that the model did not train toward the loss. I have been tuning the loss function and not getting a good result. Wonder is there any recommendation for tuning the loss function setup. Many Thanks.