xlm-roberta-base-afr-finetuned-augmentation-LUNAR

This model is a fine-tuned version of FacebookAI/xlm-roberta-base on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.2691
• F1: 0.6374
• Roc Auc: 0.7739
• Accuracy: 0.6403

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: cosine
• lr_scheduler_warmup_steps: 100
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	F1	Roc Auc	Accuracy
0.4503	1.0	76	0.3340	0.0	0.5	0.3894
0.3499	2.0	152	0.3279	0.0	0.5	0.3894
0.3336	3.0	228	0.3254	0.0	0.5	0.3894
0.3184	4.0	304	0.3118	0.1136	0.5288	0.4554
0.2854	5.0	380	0.2754	0.2345	0.5740	0.4950
0.2452	6.0	456	0.2629	0.2101	0.5669	0.5479
0.2217	7.0	532	0.2838	0.2334	0.5757	0.5248
0.1786	8.0	608	0.2559	0.2849	0.6069	0.6040
0.1762	9.0	684	0.2598	0.2635	0.5939	0.5974
0.1555	10.0	760	0.2608	0.3552	0.6270	0.6106
0.1236	11.0	836	0.2658	0.3511	0.6273	0.6073
0.1121	12.0	912	0.2603	0.3519	0.6295	0.6337
0.1071	13.0	988	0.2730	0.4093	0.6567	0.6304
0.0962	14.0	1064	0.2650	0.4117	0.6564	0.6271
0.0821	15.0	1140	0.2643	0.6313	0.7675	0.6403
0.07	16.0	1216	0.2723	0.6134	0.7644	0.6238
0.0709	17.0	1292	0.2736	0.6175	0.7648	0.6271
0.0645	18.0	1368	0.2703	0.6340	0.7715	0.6337
0.0729	19.0	1444	0.2691	0.6374	0.7739	0.6403
0.0693	20.0	1520	0.2693	0.6360	0.7735	0.6370

Framework versions

• Transformers 4.45.1
• Pytorch 2.4.0
• Datasets 3.0.1
• Tokenizers 0.20.0