9034a1fcfcb8a5ff7cbf175793008c18

This model is a fine-tuned version of google/long-t5-tglobal-xl on the Helsinki-NLP/opus_books [en-pl] dataset. It achieves the following results on the evaluation set:

• Loss: 2.1059
• Data Size: 1.0
• Epoch Runtime: 50.8968
• Bleu: 1.0080

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	3.5166	0	3.6975	0.0918
No log	1	70	3.1484	0.0078	4.5885	0.0898
No log	2	140	2.9234	0.0156	8.0484	0.1623
No log	3	210	2.8465	0.0312	12.6866	0.2145
No log	4	280	2.7703	0.0625	17.0154	0.2403
No log	5	350	2.6382	0.125	23.6663	0.3249
No log	6	420	2.5299	0.25	25.4629	0.2326
0.4562	7	490	2.4076	0.5	33.7297	0.3530
2.6037	8.0	560	2.2761	1.0	54.2496	0.4265
2.4287	9.0	630	2.1983	1.0	56.3238	0.5186
2.252	10.0	700	2.1474	1.0	49.3914	0.5376
2.1446	11.0	770	2.1003	1.0	51.7802	0.6570
2.0645	12.0	840	2.0927	1.0	54.7037	0.7908
1.9442	13.0	910	2.0730	1.0	47.2833	0.7470
1.8473	14.0	980	2.0580	1.0	48.8424	0.8291
1.7615	15.0	1050	2.0606	1.0	51.2769	0.8829
1.6682	16.0	1120	2.0772	1.0	53.6205	0.9495
1.5983	17.0	1190	2.0882	1.0	49.7303	0.9912
1.4839	18.0	1260	2.1059	1.0	50.8968	1.0080

Framework versions

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