@@ -26,7 +26,7 @@ overwrite_output_dir: true
|
||||
per_device_train_batch_size: 1
|
||||
gradient_accumulation_steps: 8
|
||||
optim: paged_adamw_8bit
|
||||
learning_rate: 1.0e-4
|
||||
learning_rate: 1.0e-5
|
||||
num_train_epochs: 3.0
|
||||
lr_scheduler_type: cosine
|
||||
warmup_ratio: 0.1
|
||||
|
||||
Reference in New Issue
Block a user