iva_mt_wslot-m2m100_418M-en-pl-nomassive

This model is a fine-tuned version of facebook/m2m100_418M on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.3751
• Bleu: 61.2295
• Gen Len: 23.3774

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: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Gen Len
0.4315	1.0	2212	0.3475	57.0703	23.338
0.2976	2.0	4424	0.3295	58.3378	23.5036
0.2153	3.0	6636	0.3347	59.3015	23.5601
0.1542	4.0	8848	0.3420	59.3942	23.4442
0.1177	5.0	11060	0.3454	60.4525	23.3501
0.0858	6.0	13272	0.3529	60.6123	23.3422
0.06	7.0	15484	0.3586	60.1342	23.4478
0.0479	8.0	17696	0.3668	60.6348	23.4163
0.0371	9.0	19908	0.3728	61.3986	23.4436
0.0299	10.0	22120	0.3751	61.2295	23.3774

Framework versions

• Transformers 4.27.4
• Pytorch 2.0.0+cu118
• Datasets 2.11.0
• Tokenizers 0.13.3