8e359c8d9ab1bae55dc31274b51dbee5

This model is a fine-tuned version of facebook/mbart-large-50 on the Helsinki-NLP/opus_books [fr-pt] dataset. It achieves the following results on the evaluation set:

• Loss: 5.9739
• Data Size: 1.0
• Epoch Runtime: 15.0577
• Bleu: 0.0597

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: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 32
• total_eval_batch_size: 32
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Data Size	Epoch Runtime	Bleu
No log	0	0	8.7118	0	1.2789	0.4609
No log	1	31	6.7838	0.0078	1.7485	0.9858
No log	2	62	5.6470	0.0156	2.2812	2.1700
No log	3	93	5.0484	0.0312	3.3009	3.6208
No log	4	124	4.7385	0.0625	4.2686	6.6987
No log	5	155	4.4426	0.125	5.6172	7.6545
No log	6	186	4.0750	0.25	7.3159	8.9402
0.6973	7	217	2.9549	0.5	9.3617	8.8149
0.6973	8.0	248	2.4132	1.0	12.5722	7.9195
1.6787	9.0	279	11.9698	1.0	12.5663	0.0
12.4061	10.0	310	7.5763	1.0	12.6830	0.0542
12.4061	11.0	341	6.4546	1.0	13.0518	0.0117
7.2485	12.0	372	5.9739	1.0	15.0577	0.0597

Framework versions

• Transformers 4.57.0
• Pytorch 2.8.0+cu128
• Datasets 4.2.0
• Tokenizers 0.22.1