afro-xlmr-base-sun-noaug

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

• Loss: 0.3348
• F1: 0.3427
• Roc Auc: 0.6065
• Accuracy: 0.4925

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 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
0.6846	1.0	29	0.5843	0.1405	0.5	0.4472
0.5818	2.0	58	0.4117	0.1405	0.5	0.4472
0.4152	3.0	87	0.4075	0.1405	0.5	0.4472
0.4037	4.0	116	0.3993	0.1405	0.5	0.4472
0.3772	5.0	145	0.3553	0.2579	0.5785	0.4874
0.3503	6.0	174	0.3497	0.2511	0.5788	0.4874
0.3288	7.0	203	0.3626	0.2516	0.5726	0.4925
0.331	8.0	232	0.3551	0.2532	0.5978	0.4523
0.2938	9.0	261	0.3443	0.2647	0.5840	0.4925
0.2872	10.0	290	0.3438	0.2757	0.5911	0.4724
0.2911	11.0	319	0.3444	0.2925	0.5987	0.4623
0.2601	12.0	348	0.3431	0.3391	0.6240	0.4724
0.274	13.0	377	0.3499	0.2773	0.5863	0.4975
0.2375	14.0	406	0.3364	0.3225	0.6114	0.5126
0.2508	15.0	435	0.3301	0.3177	0.6042	0.4874
0.2368	16.0	464	0.3348	0.3427	0.6065	0.4925
0.243	17.0	493	0.3312	0.3180	0.6058	0.4975
0.237	18.0	522	0.3319	0.3197	0.6053	0.4975
0.2217	19.0	551	0.3327	0.3281	0.6056	0.4975
0.2295	20.0	580	0.3328	0.3281	0.6056	0.4975

Framework versions

• Transformers 4.47.0
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0