xlnet-base-cased_fold_10_binary_v1

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

• Loss: 1.7782
• F1: 0.8137

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: linear
• num_epochs: 25

Training results

Training Loss	Epoch	Step	Validation Loss	F1
No log	1.0	288	0.3796	0.8145
0.4196	2.0	576	0.4319	0.7810
0.4196	3.0	864	0.6227	0.8002
0.231	4.0	1152	0.6258	0.7941
0.231	5.0	1440	1.0692	0.7866
0.1307	6.0	1728	1.1257	0.8005
0.0756	7.0	2016	1.2283	0.8072
0.0756	8.0	2304	1.3407	0.8061
0.0486	9.0	2592	1.5232	0.8059
0.0486	10.0	2880	1.6731	0.8053
0.0339	11.0	3168	1.6536	0.8087
0.0339	12.0	3456	1.7526	0.7996
0.019	13.0	3744	1.6662	0.7909
0.0237	14.0	4032	1.6028	0.8071
0.0237	15.0	4320	1.7627	0.7964
0.0078	16.0	4608	1.6513	0.8169
0.0078	17.0	4896	1.7795	0.8039
0.015	18.0	5184	1.8669	0.7935
0.015	19.0	5472	1.6288	0.8118
0.0124	20.0	5760	1.6630	0.8104
0.004	21.0	6048	1.7418	0.8167
0.004	22.0	6336	1.7651	0.8128
0.0043	23.0	6624	1.7279	0.8163
0.0043	24.0	6912	1.8177	0.8093
0.004	25.0	7200	1.7782	0.8137

Framework versions

• Transformers 4.21.1
• Pytorch 1.12.0+cu113
• Datasets 2.4.0
• Tokenizers 0.12.1