training: | |
batch_size: 4 | |
epochs: 3 | |
learning_rate: 2e-5 | |
gradient_accumulation: 8 | |
mixed_precision: "fp16" | |
save_steps: 500 | |
training: | |
batch_size: 4 | |
epochs: 3 | |
learning_rate: 2e-5 | |
gradient_accumulation: 8 | |
mixed_precision: "fp16" | |
save_steps: 500 | |