afro-xlmr-large-76L-afr-DAPT-finetuned-20-epochs

This model is a fine-tuned version of Davlan/afro-xlmr-large-76L on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.2345
• F1: 0.8736
• Roc Auc: 0.9183
• Accuracy: 0.7733

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: 32
• eval_batch_size: 32
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH 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: 100
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	F1	Roc Auc	Accuracy
0.5348	1.0	81	0.4887	0.1383	0.5513	0.1708
0.3047	2.0	162	0.2438	0.8299	0.8772	0.7034
0.1832	3.0	243	0.2173	0.8636	0.9047	0.7143
0.1559	4.0	324	0.2018	0.8632	0.9072	0.7360
0.1441	5.0	405	0.2017	0.8600	0.9012	0.7422
0.1193	6.0	486	0.2044	0.8590	0.9017	0.7407
0.0899	7.0	567	0.2345	0.8736	0.9183	0.7733
0.0793	8.0	648	0.2243	0.8688	0.9142	0.7484
0.0593	9.0	729	0.2439	0.8500	0.8945	0.7174
0.0524	10.0	810	0.2493	0.8610	0.9054	0.7422
0.054	11.0	891	0.2408	0.8672	0.9087	0.7578

Framework versions

• Transformers 4.48.1
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0