Open ydteng opened 11 months ago
I tried to reproduce it, but it had been running for a day on a Tesla K80 graphics card, and progress was extremely slow. The dataset is less than 1GB.
I am not sure how much time the training takes in K80. Even if the model is trained on 4 v100, it also needs to take about one day.
I tried to reproduce it, but it had been running for a day on a Tesla K80 graphics card, and progress was extremely slow. The dataset is less than 1GB.