le-fine-tune-mt5-small

This model is a fine-tuned version of google/mt5-small on the ravkuk_summerize_dataset dataset. It achieves the following results on the evaluation set:

• Loss: 2.4588
• Rouge1: 0.16
• Rouge2: 0.0655
• Rougel: 0.1519
• Rougelsum: 0.1522
• Gen Len: 18.9886

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: 0.0014142135623730952
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.3
• num_epochs: 8

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
7.6563	1.0	197	3.0371	0.1002	0.028	0.0941	0.0941	18.8949
3.4565	2.0	394	2.8403	0.1204	0.0362	0.1154	0.1152	18.9744
3.0924	3.0	591	2.7173	0.1276	0.0431	0.1199	0.1202	18.9631
2.7357	4.0	788	2.5831	0.1494	0.0555	0.1415	0.1416	18.9744
2.4543	5.0	985	2.5135	0.1437	0.0545	0.1345	0.135	18.9886
2.2055	6.0	1182	2.5031	0.1544	0.0642	0.147	0.1472	18.9773
2.0147	7.0	1379	2.4554	0.158	0.0643	0.1484	0.1487	18.9688
1.8746	8.0	1576	2.4588	0.16	0.0655	0.1519	0.1522	18.9886

Framework versions

• Transformers 4.26.1
• Pytorch 1.13.1+cu116
• Datasets 2.10.1
• Tokenizers 0.13.2