xlnet-large-cased-finetuned-augmentation-LUNAR

This model is a fine-tuned version of xlnet/xlnet-large-cased on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.5861
• F1: 0.7939
• Roc Auc: 0.8377
• Accuracy: 0.5610

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
0.5702	1.0	179	0.5158	0.2278	0.5480	0.2006
0.4202	2.0	358	0.3750	0.6839	0.7559	0.4376
0.336	3.0	537	0.3422	0.7237	0.7859	0.4642
0.2321	4.0	716	0.3447	0.7519	0.8097	0.5245
0.1487	5.0	895	0.3742	0.7534	0.8108	0.5316
0.0974	6.0	1074	0.4043	0.7727	0.8325	0.5442
0.0674	7.0	1253	0.4612	0.7701	0.8214	0.5175
0.0485	8.0	1432	0.4862	0.7771	0.8227	0.5330
0.0281	9.0	1611	0.5346	0.7712	0.8249	0.5456
0.0178	10.0	1790	0.5535	0.7709	0.8213	0.5372
0.0137	11.0	1969	0.5715	0.7908	0.8450	0.5484
0.0144	12.0	2148	0.5597	0.7866	0.8343	0.5694
0.005	13.0	2327	0.5850	0.7844	0.8333	0.5596
0.0034	14.0	2506	0.5807	0.7881	0.8344	0.5596
0.0034	15.0	2685	0.5856	0.7924	0.8370	0.5750
0.0025	16.0	2864	0.5861	0.7939	0.8377	0.5610
0.0056	17.0	3043	0.5916	0.7920	0.8374	0.5596
0.0043	18.0	3222	0.5900	0.7909	0.8374	0.5708
0.0036	19.0	3401	0.5902	0.7882	0.8352	0.5652
0.0016	20.0	3580	0.5903	0.7890	0.8359	0.5666

Framework versions

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