IndoT5-LoRA-R32-LR0.0001-Ep20

This model is a fine-tuned version of Wikidepia/IndoT5-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.4501

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: 0.0001
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss
0.938	1.0	188	0.8712
0.8076	2.0	376	0.8285
0.8173	3.0	564	0.7680
0.5857	4.0	752	0.5895
0.6269	5.0	940	0.5487
0.5643	6.0	1128	0.5347
0.5949	7.0	1316	0.5417
0.5524	8.0	1504	0.5260
0.5285	9.0	1692	0.5242
0.5194	10.0	1880	0.5219
0.5065	11.0	2068	0.5216
0.5097	12.0	2256	0.5141
0.5411	13.0	2444	0.5144
0.5053	14.0	2632	0.5093
0.5179	15.0	2820	0.5076
0.4767	16.0	3008	0.5031
0.4467	17.0	3196	0.4571
0.4902	18.0	3384	0.4518
0.4219	19.0	3572	0.4499
0.4428	20.0	3760	0.4501

Framework versions

• PEFT 0.18.1
• Transformers 4.57.6
• Pytorch 2.10.0+cu128
• Datasets 4.5.0
• Tokenizers 0.22.2