Open HasilPark opened 2 years ago
When the total training data can be divisible by batch_size, the model of the 20th epoch cannot be saved. This is a small problem in our code, which will be optimized later. You can save the model of the 20th epoch by adjusting the batch_size.
I trained by setting the epoch parameter to 20 using got-10k dataset. However, only epochs 1 to 19 are saved. What's the problem?