Open chenhao324 opened 7 months ago
Hello author, I used the pytorch version of the code for fine-tuning. Why is the result obtained by the first epoch significantly better than that obtained by the second epoch
learning rate
My data is over a million, and I also trained using pre trained weight parameters. I will try to modify the learning rate. Thank you for your answer
Hello author, I used the pytorch version of the code for fine-tuning. Why is the result obtained by the first epoch significantly better than that obtained by the second epoch