cartesinus/iva_mt-leyzer-intent_baseline-xlm_r-pl

fedcsis_translated-intent_baseline-xlm_r-pl

This model is a fine-tuned version of xlm-roberta-base on the leyzer-fedcsis-translated dataset.

Results on untranslated test set:

• Accuracy: 0.8769

It achieves the following results on the evaluation set:

• Loss: 0.5478
• Accuracy: 0.8769
• F1: 0.8769

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

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1
3.505	1.0	814	1.8819	0.5979	0.5979
1.5056	2.0	1628	1.1033	0.7611	0.7611
1.0892	3.0	2442	0.7402	0.8470	0.8470
0.648	4.0	3256	0.5263	0.8902	0.8902
0.423	5.0	4070	0.4253	0.9152	0.9152
0.3429	6.0	4884	0.3654	0.9194	0.9194
0.2464	7.0	5698	0.3213	0.9273	0.9273
0.1873	8.0	6512	0.3065	0.9328	0.9328
0.1666	9.0	7326	0.3046	0.9345	0.9345
0.1459	10.0	8140	0.2911	0.9370	0.9370

Framework versions

• Transformers 4.27.3
• Pytorch 1.13.1+cu116
• Datasets 2.10.1
• Tokenizers 0.13.2