learning_rate: 5e-05
train_batch_size: 12
eval_batch_size: 12
seed: 42
gradient_accumulation_steps: 32
total_train_batch_size: 384
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED 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	4.0805	2.9231	0.0	2.6635	2.6442	20.0
No log	2.0	2	3.7128	12.6859	4.5992	11.8258	12.5583	19.35
No log	3.0	3	3.4877	16.8517	8.4744	16.038	16.5316	18.35
No log	4.0	4	3.3881	18.7157	7.4343	16.8165	17.7776	16.9
No log	5.0	5	3.3881	18.7157	7.4343	16.8165	17.7776	16.9
No log	6.0	6	3.3675	20.9949	8.5536	19.7058	20.0583	15.1
No log	7.0	7	3.3574	20.5585	8.5854	18.6151	19.2469	15.25
No log	8.0	8	3.3539	21.4629	8.781	19.3981	20.0634	14.8

HuyTran1301/constrative_faiss_v2

constrative_faiss_v2