md_mt5_base_boun_split_second_v1_retrain_on_second_imst

This model is a fine-tuned version of Buseak/md_mt5_base_boun_split_second_v1_retrain_on_first_boun on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1581
• Bleu: 2.0954
• Gen Len: 18.7592

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 15

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Gen Len
0.5133	1.0	916	0.2581	1.8828	18.7472
0.4699	2.0	1832	0.2371	1.9073	18.7142
0.4469	3.0	2748	0.2222	1.9197	18.739
0.4205	4.0	3664	0.2085	1.9502	18.7502
0.4001	5.0	4580	0.1983	1.9643	18.7409
0.3992	6.0	5496	0.1898	1.9948	18.7633
0.3824	7.0	6412	0.1841	2.0136	18.7614
0.3666	8.0	7328	0.1766	2.0422	18.7649
0.3562	9.0	8244	0.1713	2.0554	18.7636
0.345	10.0	9160	0.1672	2.0659	18.7611
0.3465	11.0	10076	0.1637	2.0804	18.7611
0.3303	12.0	10992	0.1617	2.0815	18.7619
0.3333	13.0	11908	0.1594	2.0835	18.7592
0.328	14.0	12824	0.1581	2.0958	18.7592
0.3267	15.0	13740	0.1581	2.0954	18.7592

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu118
• Datasets 2.15.0
• Tokenizers 0.15.0