You can deploy this model and start earning money today!

JuliusFx/
dyu-fr-helsinki_v2_100

Name: dyu-fr-helsinki_v2_100
Availability: InStock

safetensors

marian

generated_from_trainer

Helsinki-NLP/opus-mt-en-fr

finetune

apache-2.0

Views

No views yet

dyu-fr-helsinki_v2_100

This model is a fine-tuned version of Helsinki-NLP/opus-mt-en-fr on the None dataset. It achieves the following results on the evaluation set:

Loss: 6.1859
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
4.0305	1.0	596	3.6167	0.0
3.4437	2.0	1192	3.5116	0.0
3.1234	3.0	1788	3.4381	0.0
2.8349	4.0	2384	3.4768	0.0
2.6193	5.0	2980	3.5666	0.0442
2.1941	6.0	3576	3.6278	0.0
2.0056	7.0	4172	3.6280	0.0483
1.8589	8.0	4768	3.7428	0.0486
1.7333	9.0	5364	3.7805	0.0
1.6008	10.0	5960	3.8915	0.0
1.3622	11.0	6556	3.8217	0.0615
1.2357	12.0	7152	4.0128	0.0
1.1476	13.0	7748	4.0078	0.0
1.0478	14.0	8344	4.1311	0.0
0.9838	15.0	8940	4.1934	0.0
0.8181	16.0	9536	4.1675	0.0
0.7396	17.0	10132	4.2280	0.0
0.6832	18.0	10728	4.4568	0.0
0.6346	19.0	11324	4.4663	0.0
0.5753	20.0	11920	4.5849	0.0
0.4835	21.0	12516	4.7005	0.0
0.4304	22.0	13112	4.6964	0.0
0.3909	23.0	13708	4.9668	0.0
0.362	24.0	14304	4.8289	0.0
0.3336	25.0	14900	4.9156	0.0
0.309	26.0	15496	4.8652	0.0504
0.251	27.0	16092	4.9964	0.0602
0.2299	28.0	16688	5.0098	0.0
0.2107	29.0	17284	5.0856	0.0473
0.1986	30.0	17880	5.2101	0.0
0.1852	31.0	18476	5.2712	0.0433
0.1521	32.0	19072	5.1357	0.0569
0.1396	33.0	19668	5.1737	0.0477
0.1286	34.0	20264	5.3869	0.0
0.1247	35.0	20860	5.3110	0.0569
0.1143	36.0	21456	5.4105	0.0586
0.0995	37.0	22052	5.1111	0.0
0.092	38.0	22648	5.3818	0.0
0.0892	39.0	23244	5.3847	0.0545
0.0831	40.0	23840	5.5160	0.0
0.0785	41.0	24436	5.4548	0.0
0.0709	42.0	25032	5.4513	0.0
0.0659	43.0	25628	5.4427	0.0
0.0618	44.0	26224	5.4374	0.0
0.0589	45.0	26820	5.4812	0.0
0.0578	46.0	27416	5.4300	0.0
0.0523	47.0	28012	5.5686	0.0
0.0483	48.0	28608	5.5325	0.0
0.0464	49.0	29204	5.5621	0.0
0.0445	50.0	29800	5.6814	0.0
0.0442	51.0	30396	5.7346	0.0
0.0403	52.0	30992	5.7778	0.0
0.0369	53.0	31588	5.8471	0.0
0.0356	54.0	32184	5.7581	0.0
0.0332	55.0	32780	5.8160	0.0
0.0339	56.0	33376	5.7424	0.0
0.0329	57.0	33972	5.7840	0.0
0.0297	58.0	34568	5.8663	0.0
0.0297	59.0	35164	5.7769	0.0
0.0286	60.0	35760	5.9370	0.0
0.0274	61.0	36356	5.8307	0.0
0.0258	62.0	36952	5.9233	0.0
0.0237	63.0	37548	5.9317	0.0
0.0229	64.0	38144	6.0147	0.0
0.0231	65.0	38740	6.0108	0.0
0.022	66.0	39336	6.0008	0.0
0.0218	67.0	39932	6.1052	0.0
0.0207	68.0	40528	6.0856	0.0
0.0196	69.0	41124	6.0324	0.0
0.0189	70.0	41720	5.9874	0.0
0.0191	71.0	42316	5.9704	0.0
0.018	72.0	42912	6.0590	0.0
0.017	73.0	43508	6.0278	0.0
0.0161	74.0	44104	5.8944	0.0
0.0165	75.0	44700	6.0230	0.0
0.0165	76.0	45296	5.9187	0.0
0.0151	77.0	45892	5.9900	0.0
0.0152	78.0	46488	5.9495	0.0
0.0142	79.0	47084	6.0193	0.0
0.0136	80.0	47680	6.0593	0.0
0.015	81.0	48276	6.1043	0.0
0.0138	82.0	48872	6.1035	0.0
0.013	83.0	49468	6.0649	0.0
0.0127	84.0	50064	6.1442	0.0
0.0118	85.0	50660	6.2428	0.0
0.0122	86.0	51256	6.1664	0.0
0.0119	87.0	51852	6.1649	0.0
0.0124	88.0	52448	6.1618	0.0
0.0114	89.0	53044	6.1934	0.0
0.0113	90.0	53640	6.1823	0.0
0.0104	91.0	54236	6.1765	0.0
0.0112	92.0	54832	6.2050	0.0
0.0107	93.0	55428	6.1729	0.0
0.0106	94.0	56024	6.1775	0.0
0.0107	95.0	56620	6.1426	0.0
0.0096	96.0	57216	6.1323	0.0
0.0099	97.0	57812	6.1661	0.0
0.0096	98.0	58408	6.1723	0.0
0.0099	99.0	59004	6.1834	0.0
0.0097	100.0	59600	6.1859	0.0

Framework versions

Transformers 4.44.0
Pytorch 2.1.0+cu118
Datasets 2.20.0
Tokenizers 0.19.1

JuliusFx/dyu-fr-helsinki_v2_100

dyu-fr-helsinki_v2_100

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

JuliusFx/
dyu-fr-helsinki_v2_100