mBART_translator_json_2

This model is a fine-tuned version of facebook/mbart-large-cc25 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.1203
• Bleu: 77.8658
• Gen Len: 36.1527

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: 8
• eval_batch_size: 8
• 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
1.7858	1.0	1912	0.6568	55.2937	75.6389
0.994	2.0	3824	0.4015	71.3655	35.744
0.7267	3.0	5736	0.2971	66.7522	34.5473
0.5916	4.0	7648	0.2437	80.0233	37.4331
0.502	5.0	9560	0.2072	80.9632	36.9833
0.433	6.0	11472	0.1767	69.9384	36.6381
0.3581	7.0	13384	0.1566	64.615	34.8954
0.3244	8.0	15296	0.1382	77.5563	36.1736
0.2815	9.0	17208	0.1259	76.1662	36.1548
0.2555	10.0	19120	0.1203	77.8658	36.1527

Framework versions

• Transformers 4.23.1
• Pytorch 1.12.1+cu113
• Datasets 2.5.2
• Tokenizers 0.13.1