92750bf2de129aeb373ebbbb3b09c01e

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

• Loss: 1.0092
• Data Size: 1.0
• Epoch Runtime: 238.5313
• Bleu: 13.4589

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	2.8029	0	18.0919	0.1281
No log	1	437	2.3923	0.0078	21.2322	0.2975
No log	2	874	2.2184	0.0156	24.8742	0.3557
No log	3	1311	2.0426	0.0312	32.5722	0.7512
No log	4	1748	1.8893	0.0625	41.3585	1.4739
2.0564	5	2185	1.7338	0.125	57.0915	2.7085
1.8861	6	2622	1.5660	0.25	80.9126	4.4450
1.6364	7	3059	1.3907	0.5	135.0631	5.7552
1.3652	8.0	3496	1.1953	1.0	245.6774	7.9893
1.1932	9.0	3933	1.0847	1.0	237.4980	9.5079
1.0438	10.0	4370	1.0209	1.0	240.7625	10.4328
0.9525	11.0	4807	0.9845	1.0	238.4553	11.1651
0.8678	12.0	5244	0.9473	1.0	238.6869	11.9281
0.787	13.0	5681	0.9370	1.0	238.1721	12.3069
0.7035	14.0	6118	0.9371	1.0	238.8203	12.7342
0.6576	15.0	6555	0.9317	1.0	237.5394	12.9370
0.5825	16.0	6992	0.9412	1.0	238.2312	13.3106
0.537	17.0	7429	0.9595	1.0	237.2018	13.3257
0.4899	18.0	7866	0.9831	1.0	236.6620	13.5000
0.4346	19.0	8303	1.0092	1.0	238.5313	13.4589

Framework versions

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