opus-mt-tc-big-hu-en-finetuned-news

This model is a fine-tuned version of Helsinki-NLP/opus-mt-tc-big-hu-en on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.1412
• Bleu: 40.3642
• Gen Len: 44.0303

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: 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: 3

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Gen Len
No log	0.11	100	1.1822	38.1953	43.5551
No log	0.22	200	1.1738	38.6712	43.549
No log	0.32	300	1.1602	39.2014	44.2009
No log	0.43	400	1.1503	39.2141	43.9468
1.2199	0.54	500	1.1451	39.2687	43.6871
1.2199	0.65	600	1.1349	39.445	43.9483
1.2199	0.75	700	1.1356	39.3787	43.47
1.2199	0.86	800	1.1233	39.7025	43.9054
1.2199	0.97	900	1.1224	39.9764	43.9656
1.1308	1.08	1000	1.1343	39.8533	43.9929
1.1308	1.19	1100	1.1446	39.7232	43.675
1.1308	1.29	1200	1.1378	40.0687	44.0606
1.1308	1.4	1300	1.1324	39.9239	43.7738
1.1308	1.51	1400	1.1330	40.0318	43.7756
0.8661	1.62	1500	1.1315	39.8677	43.7542
0.8661	1.72	1600	1.1185	40.1978	44.168
0.8661	1.83	1700	1.1298	40.254	44.0497
0.8661	1.94	1800	1.1191	40.2197	44.0295
0.8661	2.05	1900	1.1416	40.1255	44.0534
0.8198	2.16	2000	1.1479	40.3099	43.9854
0.8198	2.26	2100	1.1495	40.3473	44.0204
0.8198	2.37	2200	1.1453	40.329	44.0764
0.8198	2.48	2300	1.1450	40.2623	44.0944
0.8198	2.59	2400	1.1471	40.416	44.1797
0.6783	2.69	2500	1.1433	40.4645	44.0817
0.6783	2.8	2600	1.1405	40.4229	44.0554
0.6783	2.91	2700	1.1418	40.4142	44.0493

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu121
• Datasets 2.17.0
• Tokenizers 0.15.1