xlnet-base-cased_fold_8_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: 1.5333
• F1: 0.8407

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	290	0.3866	0.8172
0.4299	2.0	580	0.4215	0.8246
0.4299	3.0	870	0.4765	0.8238
0.2564	4.0	1160	0.7283	0.8350
0.2564	5.0	1450	0.6825	0.8363
0.1553	6.0	1740	0.9637	0.8339
0.0893	7.0	2030	1.1392	0.8239
0.0893	8.0	2320	1.1868	0.8231
0.0538	9.0	2610	1.2180	0.8346
0.0538	10.0	2900	1.2353	0.8253
0.0386	11.0	3190	1.1883	0.8317
0.0386	12.0	3480	1.2786	0.8375
0.0289	13.0	3770	1.3725	0.8375
0.0146	14.0	4060	1.3171	0.8463
0.0146	15.0	4350	1.2323	0.8425
0.0182	16.0	4640	1.3169	0.8485
0.0182	17.0	4930	1.4424	0.8336
0.0125	18.0	5220	1.4336	0.8385
0.0102	19.0	5510	1.4888	0.8405
0.0102	20.0	5800	1.5227	0.8419
0.0035	21.0	6090	1.4994	0.8421
0.0035	22.0	6380	1.4845	0.8424
0.0047	23.0	6670	1.5006	0.8422
0.0047	24.0	6960	1.5468	0.8422
0.0042	25.0	7250	1.5333	0.8407

Framework versions

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