Visualize in Weights & Biases

Bamalingua-11

This model is a fine-tuned version of facebook/mbart-large-50-many-to-many-mmt on an unknown dataset. It achieves the following results on the evaluation set:

  • Loss: 1.7913
  • Bleu: 33.0834

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: 0.0001
  • train_batch_size: 8
  • eval_batch_size: 8
  • seed: 42
  • gradient_accumulation_steps: 2
  • total_train_batch_size: 16
  • optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
  • lr_scheduler_type: cosine
  • lr_scheduler_warmup_steps: 500
  • num_epochs: 9

Training results

Training Loss Epoch Step Validation Loss Bleu
2.0735 1.0 1580 1.7203 13.4475
1.2602 2.0 3160 1.4606 19.4716
0.7678 3.0 4740 1.4131 25.8137
0.4892 4.0 6320 1.4349 29.8337
0.3242 5.0 7900 1.4621 32.3790
0.2209 6.0 9480 1.4949 32.8099
0.1498 7.0 11060 1.5509 32.9551
0.0964 8.0 12640 1.6539 33.0301
0.0595 9.0 14220 1.7913 33.0834

Framework versions

  • Transformers 4.57.1
  • Pytorch 2.8.0+cu126
  • Datasets 4.4.2
  • Tokenizers 0.22.1
Downloads last month
3
Safetensors
Model size
0.6B params
Tensor type
F32
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support

Model tree for Gaoussin/Bamalingua-11

Finetuned
(221)
this model