CS221-xlm-roberta-base-finetuned-amh

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

• Loss: 0.5618
• F1: 0.6083
• Roc Auc: 0.7556
• Accuracy: 0.4645

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: 8
• eval_batch_size: 8
• 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
0.4096	1.0	444	0.4063	0.1607	0.5416	0.2213
0.3316	2.0	888	0.3469	0.3841	0.6362	0.3649
0.2866	3.0	1332	0.3261	0.4003	0.6436	0.3902
0.2372	4.0	1776	0.3272	0.4436	0.6579	0.4037
0.2571	5.0	2220	0.3277	0.4953	0.6924	0.4240
0.1821	6.0	2664	0.3426	0.5072	0.7043	0.4274
0.1917	7.0	3108	0.3656	0.4951	0.6977	0.4172
0.1775	8.0	3552	0.3620	0.5292	0.7094	0.4426
0.1457	9.0	3996	0.3958	0.5770	0.7342	0.4443
0.1026	10.0	4440	0.4400	0.5745	0.7357	0.4223
0.0671	11.0	4884	0.4693	0.5869	0.7383	0.4307
0.0587	12.0	5328	0.4951	0.5807	0.7371	0.4291
0.0548	13.0	5772	0.5063	0.6018	0.7557	0.4611
0.0654	14.0	6216	0.5344	0.5991	0.7487	0.4510
0.0613	15.0	6660	0.5403	0.6021	0.7527	0.4459
0.0471	16.0	7104	0.5473	0.6047	0.7490	0.4595
0.0169	17.0	7548	0.5571	0.6056	0.7546	0.4476
0.0178	18.0	7992	0.5618	0.6083	0.7556	0.4645
0.0317	19.0	8436	0.5584	0.6032	0.7511	0.4628
0.0325	20.0	8880	0.5587	0.6064	0.7537	0.4628

Framework versions

• Transformers 4.47.1
• Pytorch 2.5.1+cu121
• Datasets 3.2.0
• Tokenizers 0.21.0