zeon8985army/IndonesiaLukasLargeV3-2

This model is a fine-tuned version of zeon8985army/IndonesiaLukasLargeV3-2 on the Mozilla & GoogleFleur dataset. It achieves the following results on the evaluation set:

• Loss: 0.1613

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: 6e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 12
• training_steps: 552
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.2396	0.0785	23	0.2246
0.2199	0.1570	46	0.2234
0.1963	0.2355	69	0.2176
0.206	0.3140	92	0.2117
0.1755	0.3925	115	0.1899
0.1713	0.4710	138	0.2041
0.2126	0.5495	161	0.1806
0.1778	0.6280	184	0.1787
0.1646	0.7065	207	0.1767
0.1674	0.7850	230	0.1746
0.1545	0.8635	253	0.1725
0.1742	0.9420	276	0.1708
0.1434	1.0205	299	0.1687
0.15	1.0990	322	0.1679
0.1374	1.1775	345	0.1671
0.136	1.2560	368	0.1650
0.1407	1.3345	391	0.1642
0.153	1.4130	414	0.1629
0.1505	1.4915	437	0.1632
0.1543	1.5700	460	0.1624
0.1394	1.6485	483	0.1616
0.1255	1.7270	506	0.1617
0.1342	1.8055	529	0.1613
0.1459	1.8840	552	0.1613

Framework versions

• PEFT 0.9.0
• Transformers 4.44.2
• Pytorch 2.4.1+cu121
• Datasets 3.0.1
• Tokenizers 0.19.1