combined_marianmt-sanlish-to-bangla

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

Model description

More information needed

More information needed

More information needed

The following hyperparameters were used during training:

learning_rate: 2e-05
train_batch_size: 8
eval_batch_size: 8
seed: 42
optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
lr_scheduler_warmup_ratio: 0.1
num_epochs: 25
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Bleu	Chrf	Meteor	Bertscore
2.7286	1.0	246	2.5340	0.2965	10.2002	0.0332	0.7248
2.3221	2.0	492	2.1338	1.3999	15.2773	0.0690	0.7533
1.9794	3.0	738	1.8753	2.1629	21.5978	0.1526	0.7884
1.6741	4.0	984	1.7109	6.1230	27.0032	0.2034	0.8084
1.478	5.0	1230	1.5948	8.0855	32.3351	0.2460	0.8201
1.3367	6.0	1476	1.5395	8.8875	33.8923	0.2660	0.8266
1.1612	7.0	1722	1.4789	11.0814	36.0270	0.2916	0.8375
0.9993	8.0	1968	1.4588	9.8384	36.1198	0.2942	0.8400
0.9123	9.0	2214	1.4346	12.6195	37.9126	0.3224	0.8465
0.8282	10.0	2460	1.4372	12.4234	38.9587	0.3179	0.8467
0.7706	11.0	2706	1.4185	12.9207	40.1283	0.3317	0.8526
0.6913	12.0	2952	1.4071	12.8816	39.5920	0.3371	0.8502
0.6156	13.0	3198	1.4139	13.4318	41.0966	0.3399	0.8558
0.5456	14.0	3444	1.4160	13.4640	40.8981	0.3430	0.8517
0.5337	15.0	3690	1.4159	14.6363	41.3339	0.3532	0.8541
0.4909	16.0	3936	1.4243	14.8354	43.0015	0.3674	0.8635
0.4375	17.0	4182	1.4227	15.8494	42.6831	0.3610	0.8595
0.4019	18.0	4428	1.4398	15.4707	43.1817	0.3670	0.8624
0.3436	19.0	4674	1.4435	16.9224	43.5817	0.3721	0.8606
0.3378	20.0	4920	1.4413	15.0569	43.0864	0.3631	0.8633
0.3187	21.0	5166	1.4491	15.4800	43.5444	0.3664	0.8619
0.3116	22.0	5412	1.4516	15.1993	43.4683	0.3620	0.8630
0.3102	23.0	5658	1.4478	15.1517	43.6839	0.3688	0.8625
0.3273	24.0	5904	1.4504	14.8275	43.9419	0.3737	0.8636