t5-base-finetune-thai-to-romanized

This model is a fine-tuned version of t5-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 5.5823
• Rouge1: 0.0533
• Rouge2: 0.0
• Rougel: 0.0543
• Rougelsum: 0.0541
• Gen Len: 11.3346

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Rouge1	Rouge2	Rougel	Rougelsum	Gen Len
6.2299	1.0	1500	6.0909	0.0	0.0	0.0	0.0	12.4211
6.0799	2.0	3000	5.9716	0.0058	0.0	0.006	0.0062	11.1185
5.9595	3.0	4500	5.9016	0.0081	0.0	0.0082	0.008	12.1465
5.8423	4.0	6000	5.7665	0.0364	0.0017	0.0367	0.0365	11.6982
5.7915	5.0	7500	5.7470	0.0441	0.0013	0.0445	0.0442	10.7897
5.7114	6.0	9000	5.6844	0.0472	0.0013	0.0476	0.0478	10.5441
5.6555	7.0	10500	5.6416	0.0549	0.0013	0.0552	0.0552	11.1098
5.6362	8.0	12000	5.6036	0.0493	0.0013	0.0501	0.0498	12.7922
5.5941	9.0	13500	5.5895	0.0535	0.0	0.0544	0.0542	11.5989
5.569	10.0	15000	5.5823	0.0533	0.0	0.0543	0.0541	11.3346

Framework versions

• Transformers 4.26.1
• Pytorch 1.13.1+cu116
• Datasets 2.9.0
• Tokenizers 0.13.2