dyu-fr-helsinki-no-de

This model is a fine-tuned version of Helsinki-NLP/opus-mt-no-de on the None dataset. It achieves the following results on the evaluation set:

• Loss: 3.5209
• Bleu score: 0.0

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: 2e-05
• train_batch_size: 16
• eval_batch_size: 20
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu score
2.4369	1.0	596	2.6018	0.0
2.2688	2.0	1192	2.5322	0.0
2.1603	3.0	1788	2.4548	0.0
2.1341	4.0	2384	2.4211	0.0556
2.1043	5.0	2980	2.3665	0.0
1.9387	6.0	3576	2.3617	0.0
1.8674	7.0	4172	2.3577	0.0
1.8055	8.0	4768	2.3144	0.0
1.7479	9.0	5364	2.2869	0.0
1.6936	10.0	5960	2.3116	0.0639
1.5804	11.0	6556	2.3005	0.0670
1.5239	12.0	7152	2.3582	0.0646
1.4802	13.0	7748	2.3175	0.0
1.4258	14.0	8344	2.3241	0.0730
1.3863	15.0	8940	2.3482	0.0617
1.2985	16.0	9536	2.3589	0.0
1.2504	17.0	10132	2.3422	0.0579
1.216	18.0	10728	2.4315	0.0649
1.1855	19.0	11324	2.4391	0.0
1.1482	20.0	11920	2.4343	0.0698
1.0754	21.0	12516	2.4569	0.0
1.0266	22.0	13112	2.4776	0.0
1.0047	23.0	13708	2.5027	0.0
0.9649	24.0	14304	2.5325	0.0688
0.9472	25.0	14900	2.5435	0.0
0.911	26.0	15496	2.5464	0.0
0.8457	27.0	16092	2.6037	0.0597
0.8214	28.0	16688	2.6060	0.0622
0.7928	29.0	17284	2.6419	0.0654
0.7723	30.0	17880	2.6794	0.0
0.7467	31.0	18476	2.6646	0.0
0.695	32.0	19072	2.6633	0.0687
0.6656	33.0	19668	2.6997	0.0703
0.6532	34.0	20264	2.7460	0.0
0.6345	35.0	20860	2.7005	0.0
0.6048	36.0	21456	2.7600	0.0
0.5693	37.0	22052	2.7855	0.0571
0.5433	38.0	22648	2.8036	0.0
0.5311	39.0	23244	2.8011	0.0
0.5157	40.0	23840	2.8985	0.0
0.4959	41.0	24436	2.8801	0.0637
0.4689	42.0	25032	2.9200	0.0
0.4493	43.0	25628	2.9296	0.0670
0.4347	44.0	26224	2.9424	0.0
0.4179	45.0	26820	2.9265	0.0
0.4098	46.0	27416	2.9732	0.0
0.381	47.0	28012	2.9769	0.0706
0.3693	48.0	28608	2.9596	0.0
0.3547	49.0	29204	3.0723	0.0597
0.3522	50.0	29800	3.0764	0.0
0.3376	51.0	30396	3.0528	0.0
0.3335	52.0	30992	3.0604	0.0
0.3082	53.0	31588	3.0628	0.0
0.3016	54.0	32184	3.0979	0.0
0.2923	55.0	32780	3.1130	0.0
0.2854	56.0	33376	3.0929	0.0
0.2768	57.0	33972	3.1772	0.0
0.2603	58.0	34568	3.1767	0.0
0.2538	59.0	35164	3.1763	0.0
0.2509	60.0	35760	3.2054	0.0
0.2413	61.0	36356	3.2254	0.0
0.2362	62.0	36952	3.2636	0.0
0.2265	63.0	37548	3.2564	0.0
0.2198	64.0	38144	3.2841	0.0
0.2142	65.0	38740	3.2951	0.0
0.2106	66.0	39336	3.3296	0.0
0.2041	67.0	39932	3.2849	0.0
0.1983	68.0	40528	3.3286	0.0
0.1929	69.0	41124	3.3398	0.0
0.1883	70.0	41720	3.3635	0.0
0.1866	71.0	42316	3.3814	0.0
0.1817	72.0	42912	3.3958	0.0
0.1751	73.0	43508	3.4083	0.0
0.1712	74.0	44104	3.4030	0.0
0.1662	75.0	44700	3.3921	0.0
0.1664	76.0	45296	3.4042	0.0
0.1633	77.0	45892	3.4219	0.0
0.1633	78.0	46488	3.3983	0.0
0.1568	79.0	47084	3.4237	0.0
0.1525	80.0	47680	3.4105	0.0
0.1497	81.0	48276	3.4140	0.0
0.1492	82.0	48872	3.4283	0.0
0.1469	83.0	49468	3.4385	0.0
0.1412	84.0	50064	3.4418	0.0
0.1399	85.0	50660	3.4478	0.0
0.1399	86.0	51256	3.4751	0.0
0.1383	87.0	51852	3.4603	0.0
0.1363	88.0	52448	3.4558	0.0
0.1336	89.0	53044	3.4837	0.0
0.1316	90.0	53640	3.4838	0.0
0.1326	91.0	54236	3.5023	0.0
0.1302	92.0	54832	3.5134	0.0
0.1297	93.0	55428	3.5191	0.0
0.1276	94.0	56024	3.5025	0.0
0.1269	95.0	56620	3.5105	0.0
0.1258	96.0	57216	3.5082	0.0
0.126	97.0	57812	3.5234	0.0
0.1241	98.0	58408	3.5198	0.0
0.1244	99.0	59004	3.5203	0.0
0.1251	100.0	59600	3.5209	0.0

Framework versions

• Transformers 4.44.0
• Pytorch 2.1.0+cu118
• Datasets 2.21.0
• Tokenizers 0.19.1