63565e0ad765dc5aaca2cf50171e6034

This model is a fine-tuned version of google/mt5-large on the Helsinki-NLP/opus_books [en-pl] dataset. It achieves the following results on the evaluation set:

• Loss: 2.3352
• Data Size: 1.0
• Epoch Runtime: 34.6714
• Bleu: 2.9550

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: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 32
• total_eval_batch_size: 32
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Data Size	Epoch Runtime	Bleu
No log	0	0	19.5343	0	3.1551	0.0151
No log	1	70	18.2514	0.0078	3.6727	0.0157
No log	2	140	17.4263	0.0156	5.2436	0.0154
No log	3	210	17.5548	0.0312	8.8018	0.0174
No log	4	280	14.7051	0.0625	10.8488	0.0150
No log	5	350	6.1602	0.125	12.7341	0.0145
No log	6	420	4.3478	0.25	16.8816	0.0382
1.5083	7	490	3.4795	0.5	22.5077	0.0918
3.9016	8.0	560	2.8548	1.0	37.0708	0.2228
3.3784	9.0	630	2.5441	1.0	33.6755	1.7433
2.9002	10.0	700	2.4498	1.0	33.9003	2.0088
2.7537	11.0	770	2.3896	1.0	35.3899	2.1195
2.6236	12.0	840	2.3553	1.0	33.3430	2.0974
2.4693	13.0	910	2.3456	1.0	35.3635	2.2853
2.3528	14.0	980	2.3267	1.0	35.2489	2.3388
2.276	15.0	1050	2.3321	1.0	34.0896	2.5898
2.2117	16.0	1120	2.3306	1.0	34.3642	2.7405
2.1296	17.0	1190	2.3293	1.0	33.7178	2.6908
2.0273	18.0	1260	2.3352	1.0	34.6714	2.9550

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.2.0
• Tokenizers 0.22.1