le-fine-tune-mt5-base

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

• Loss: 2.6590
• Rouge1: 0.1555
• Rouge2: 0.065
• Rougel: 0.1489
• Rougelsum: 0.149
• Gen Len: 18.9858

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
5.0797	1.0	197	2.7316	0.1101	0.0319	0.1025	0.1024	18.9432
2.8975	2.0	394	2.6943	0.1239	0.0453	0.1207	0.1204	18.9688
2.7115	3.0	591	2.6143	0.1333	0.0505	0.1283	0.1289	18.9688
2.365	4.0	788	2.5704	0.125	0.0433	0.1201	0.1199	19.0
2.0738	5.0	985	2.5296	0.1341	0.0478	0.1284	0.1286	18.9858
1.6716	6.0	1182	2.4902	0.1451	0.0554	0.1397	0.1395	18.9886
1.2644	7.0	1379	2.5039	0.1446	0.0562	0.1407	0.1406	18.9744
0.9641	8.0	1576	2.6590	0.1555	0.065	0.1489	0.149	18.9858

Framework versions

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