xlnet-base-cased_fold_9_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: 1.7204
• F1: 0.8203

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	291	0.4045	0.8001
0.4262	2.0	582	0.3914	0.8297
0.4262	3.0	873	0.5050	0.8029
0.2488	4.0	1164	0.7681	0.8007
0.2488	5.0	1455	0.8349	0.8262
0.1483	6.0	1746	0.9045	0.8220
0.0894	7.0	2037	1.1584	0.8165
0.0894	8.0	2328	1.1818	0.8300
0.0389	9.0	2619	1.3332	0.8147
0.0389	10.0	2910	1.2373	0.8285
0.038	11.0	3201	1.3156	0.8234
0.038	12.0	3492	1.3251	0.8341
0.0211	13.0	3783	1.3144	0.8255
0.0158	14.0	4074	1.5686	0.8168
0.0158	15.0	4365	1.5382	0.8185
0.0165	16.0	4656	1.5203	0.8282
0.0165	17.0	4947	1.5352	0.8136
0.0142	18.0	5238	1.4799	0.8243
0.0062	19.0	5529	1.5030	0.8294
0.0062	20.0	5820	1.6264	0.8094
0.0078	21.0	6111	1.6949	0.8122
0.0078	22.0	6402	1.7106	0.8139
0.0043	23.0	6693	1.7234	0.8218
0.0043	24.0	6984	1.7344	0.8208
0.0028	25.0	7275	1.7204	0.8203

Framework versions

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