tokipona_to_eng_model_v0.1

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

• Loss: 5.0757
• Bleu: 2.1864
• Gen Len: 11.867

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: 0.002
• 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
• num_epochs: 15
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Gen Len
No log	1.0	55	3.7998	1.6168	13.0394
No log	2.0	110	3.6119	0.7534	13.2315
No log	3.0	165	3.6447	0.6867	13.0443
No log	4.0	220	3.7115	1.0019	12.0148
No log	5.0	275	3.8782	1.3715	13.2217
No log	6.0	330	4.0107	1.7444	11.266
No log	7.0	385	4.1611	2.7707	11.665
No log	8.0	440	4.3828	3.0123	12.0985
No log	9.0	495	4.5123	3.0296	12.6502
2.3706	10.0	550	4.6470	2.3476	11.8768
2.3706	11.0	605	4.8186	2.0611	12.1182
2.3706	12.0	660	4.8997	2.173	11.6995
2.3706	13.0	715	4.9742	2.2424	12.1576
2.3706	14.0	770	5.0570	2.0142	12.2611
2.3706	15.0	825	5.0757	2.1864	11.867

Framework versions

• Transformers 4.28.0
• Pytorch 2.0.0+cu118
• Datasets 2.11.0
• Tokenizers 0.13.3