md_mt5_2611_retrain_v20_imst

This model is a fine-tuned version of Buseak/md_mt5_1911_v18_retrain on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1318
• Bleu: 2.1555
• Gen Len: 18.766

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
0.4201	1.0	916	0.2039	1.9544	18.7368
0.396	2.0	1832	0.1862	2.0084	18.7398
0.363	3.0	2748	0.1757	2.0274	18.7475
0.3559	4.0	3664	0.1681	2.0622	18.7573
0.3389	5.0	4580	0.1606	2.0798	18.7614
0.3338	6.0	5496	0.1528	2.1003	18.7581
0.3192	7.0	6412	0.1476	2.1099	18.7622
0.3162	8.0	7328	0.1449	2.1171	18.7554
0.2958	9.0	8244	0.1405	2.1293	18.7559
0.3012	10.0	9160	0.1379	2.1451	18.7606
0.2909	11.0	10076	0.1352	2.1459	18.7644
0.2869	12.0	10992	0.1338	2.1467	18.7628
0.2875	13.0	11908	0.1324	2.1467	18.7639
0.2875	14.0	12824	0.1317	2.155	18.7649
0.2836	15.0	13740	0.1318	2.1555	18.766

Framework versions

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