md_mt5_0109_v7

This model is a fine-tuned version of Buseak/md_mt5_0109_v6 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0600
• Bleu: 0.6466
• Gen Len: 18.9431

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
0.1769	1.0	975	0.0816	0.6214	18.9392
0.1686	2.0	1950	0.0776	0.6253	18.9385
0.1668	3.0	2925	0.0742	0.6317	18.941
0.1618	4.0	3900	0.0711	0.6321	18.9413
0.1596	5.0	4875	0.0692	0.6355	18.9408
0.1575	6.0	5850	0.0670	0.6403	18.9426
0.1486	7.0	6825	0.0661	0.6419	18.9423
0.1504	8.0	7800	0.0636	0.6447	18.9428
0.1476	9.0	8775	0.0632	0.6443	18.9423
0.1432	10.0	9750	0.0623	0.6447	18.9428
0.1438	11.0	10725	0.0610	0.6464	18.9423
0.1411	12.0	11700	0.0607	0.6453	18.9431
0.1425	13.0	12675	0.0604	0.6475	18.9431
0.1437	14.0	13650	0.0602	0.6466	18.9431
0.1416	15.0	14625	0.0600	0.6466	18.9431

Framework versions

• Transformers 4.38.1
• Pytorch 2.1.0+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2