learning_rate: 2e-05
train_batch_size: 32
eval_batch_size: 32
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: linear
lr_scheduler_warmup_steps: 500
num_epochs: 10
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	F1	Precision	Recall
0.6924	1.0	173	0.6923	0.5665	0.5683	0.5717
0.6813	2.0	346	0.6619	0.6175	0.6288	0.6355
0.6277	3.0	519	0.6212	0.6624	0.6652	0.6606
0.5215	4.0	692	0.5947	0.6815	0.6831	0.6933
0.4264	5.0	865	0.5633	0.7342	0.7318	0.7384
0.3438	6.0	1038	0.6045	0.7624	0.7760	0.7556
0.2079	7.0	1211	0.6439	0.7584	0.7616	0.7559
0.2579	8.0	1384	0.6854	0.7350	0.7333	0.7445
0.1349	9.0	1557	0.7107	0.7682	0.7663	0.7706
0.1296	10.0	1730	0.7187	0.7689	0.7676	0.7704

usyk/tamil_bert_sharetask_batch32

tamil_bert_sharetask_batch32