hi_do_epoch_10

This model is a fine-tuned version of facebook/nllb-200-distilled-600M on the None dataset. It achieves the following results on the evaluation set:

• Loss: 1.1901
• Bleu: 40.8348
• Gen Len: 40.7269

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: 16
• 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
No log	1.0	398	2.0113	22.4938	38.9396
2.9232	2.0	796	1.6090	30.6633	39.8099
1.7621	3.0	1194	1.4372	34.6834	40.3153
1.4831	4.0	1592	1.3466	36.6566	40.6136
1.4831	5.0	1990	1.2808	38.344	40.7986
1.3346	6.0	2388	1.2338	39.2976	40.6419
1.233	7.0	2786	1.2122	40.1246	40.6923
1.1768	8.0	3184	1.1983	40.5123	40.7093
1.1437	9.0	3582	1.1935	40.8092	40.815
1.1437	10.0	3980	1.1901	40.8348	40.7269

Framework versions

• Transformers 4.26.0
• Pytorch 1.10.1+cu113
• Datasets 2.8.0
• Tokenizers 0.13.2