iva_mt_wslot-m2m100_418M-en-hi

This model is a fine-tuned version of facebook/m2m100_418M on the iva_mt_wslot dataset. It achieves the following results on the evaluation set:

• Loss: 0.0128
• Bleu: 66.566
• Gen Len: 21.9557

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: 7
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Bleu	Gen Len
0.0173	1.0	1676	0.0143	62.0504	21.7847
0.0116	2.0	3352	0.0127	64.7753	22.0289
0.0085	3.0	5028	0.0123	65.6406	21.8877
0.0064	4.0	6704	0.0123	66.4173	21.9532
0.0048	5.0	8380	0.0125	66.2846	21.8706
0.0037	6.0	10056	0.0127	66.4123	21.8911
0.003	7.0	11732	0.0128	66.566	21.9557

Framework versions

• Transformers 4.28.1
• Pytorch 2.0.1+cu118
• Datasets 2.12.0
• Tokenizers 0.13.3

Citation

If you use this model, please cite the following:

@article{Sowanski2023SlotLI,
  title={Slot Lost in Translation? Not Anymore: A Machine Translation Model for Virtual Assistants with Type-Independent Slot Transfer},
  author={Marcin Sowanski and Artur Janicki},
  journal={2023 30th International Conference on Systems, Signals and Image Processing (IWSSIP)},
  year={2023},
  pages={1-5}
}