learning_rate: 5e-05
train_batch_size: 10
eval_batch_size: 2
seed: 42
gradient_accumulation_steps: 32
total_train_batch_size: 320
optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
num_epochs: 8
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
No log	1.0	1	5.1699	5.6856	0.0	4.8309	5.0576	20.0
No log	2.0	2	5.1699	5.6856	0.0	4.8309	5.0576	20.0
No log	3.0	3	5.1699	5.6856	0.0	4.8309	5.0576	20.0
No log	4.0	4	4.0622	2.4	0.0	1.6	1.6	18.7
No log	5.0	5	3.5637	6.7206	1.0526	6.0343	6.0343	14.7
No log	6.0	6	3.3110	8.5233	0.0	6.575	6.9583	10.8
No log	7.0	7	3.1320	14.0635	3.0526	12.1905	12.1905	13.5
No log	8.0	8	3.0823	12.5556	3.25	10.4444	10.2222	11.8

HuyTran1301/constrative_cont

constrative_cont