learning_rate: 1e-05
train_batch_size: 32
eval_batch_size: 32
seed: 42
optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
num_epochs: 10
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Precision	Recall	F1
0.404	1.0	185	0.2996	0.6829	0.3218	0.4375
0.2827	2.0	370	0.2278	0.8144	0.6054	0.6945
0.2288	3.0	555	0.2090	0.815	0.6245	0.7072
0.1982	4.0	740	0.2250	0.7920	0.6858	0.7351
0.18	5.0	925	0.2729	0.8812	0.5402	0.6698
0.1703	6.0	1110	0.3118	0.8867	0.5096	0.6472
0.1604	7.0	1295	0.2985	0.8696	0.6130	0.7191
0.1522	8.0	1480	0.2928	0.8402	0.6245	0.7165
0.149	9.0	1665	0.2910	0.8168	0.6322	0.7127
0.1375	10.0	1850	0.3020	0.8308	0.6207	0.7105

Kenazin/xlnet-fine-tuned-matching

xlnet-fine-tuned-matching