CS221-xlnet-large-cased-finetuned-semeval

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

• Loss: 0.6440
• F1: 0.7846
• Roc Auc: 0.8368
• Accuracy: 0.4892

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: 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
0.577	1.0	139	0.5832	0.4593	0.6262	0.1516
0.5274	2.0	278	0.5171	0.4943	0.6506	0.1805
0.429	3.0	417	0.3725	0.7256	0.7927	0.4188
0.3146	4.0	556	0.3735	0.7468	0.8089	0.4567
0.2192	5.0	695	0.3846	0.7625	0.8216	0.4729
0.167	6.0	834	0.4137	0.7541	0.8126	0.4585
0.0954	7.0	973	0.4414	0.7672	0.8222	0.4783
0.0668	8.0	1112	0.5105	0.7696	0.8271	0.4747
0.0299	9.0	1251	0.5573	0.7688	0.8263	0.4531
0.0258	10.0	1390	0.5910	0.7793	0.8366	0.4783
0.0132	11.0	1529	0.6008	0.7741	0.8298	0.4801
0.0082	12.0	1668	0.6108	0.7780	0.8340	0.4711
0.0054	13.0	1807	0.6386	0.7806	0.8356	0.4711
0.0033	14.0	1946	0.6429	0.7775	0.8325	0.4747
0.0025	15.0	2085	0.6464	0.7763	0.8314	0.4675
0.0028	16.0	2224	0.6440	0.7846	0.8368	0.4892
0.0029	17.0	2363	0.6455	0.7816	0.8344	0.4856
0.0029	18.0	2502	0.6496	0.7777	0.8316	0.4765
0.0023	19.0	2641	0.6500	0.7812	0.8347	0.4819

Framework versions

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