Sinhala_Bert_Finetune_BS16_LR2e-5

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

• Loss: 1.2344
• Bleu: 0.7510

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: 32
• eval_batch_size: 32
• 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: linear
• num_epochs: 50
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu
8.0871	1.0	113	3.5867	1.0422
5.9030	2.0	226	2.8223	4.0487
4.8612	3.0	339	2.4661	4.7815
4.1543	4.0	452	2.1761	4.4447
3.4187	5.0	565	1.9373	11.1311
2.9872	6.0	678	1.7573	13.9844
2.6153	7.0	791	1.6216	15.2076
2.0958	8.0	904	1.5003	28.0665
1.8536	9.0	1017	1.4430	30.9662
1.5720	10.0	1130	1.3675	39.8024
1.3503	11.0	1243	1.3320	43.0020
1.2505	12.0	1356	1.3036	46.8129
1.0905	13.0	1469	1.2744	47.9441
0.9531	14.0	1582	1.2241	45.7459
0.8876	15.0	1695	1.2142	43.2317

Framework versions

• Transformers 5.0.0
• Pytorch 2.10.0+cu128
• Datasets 4.8.5
• Tokenizers 0.22.2