learning_rate: 2e-05
train_batch_size: 128
eval_batch_size: 128
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: 20
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Accuracy	F1 Macro
No log	1.0	44	1.0907	0.3793	0.3607
1.103	2.0	88	1.0761	0.3971	0.3968
1.0717	3.0	132	1.0490	0.4493	0.4502
1.0017	4.0	176	1.0444	0.4721	0.4709
0.9095	5.0	220	1.0969	0.4721	0.4701
0.8033	6.0	264	1.1493	0.4993	0.4979
0.6794	7.0	308	1.2285	0.5121	0.5084
0.5809	8.0	352	1.3225	0.5071	0.5035
0.5809	9.0	396	1.5689	0.4771	0.4631
0.4943	10.0	440	1.4892	0.5036	0.4998
0.424	11.0	484	1.6690	0.4836	0.4802

Kuongan/DSC_visobert_finetuned

DSC_visobert_finetuned