learning_rate: 5e-05
train_batch_size: 128
eval_batch_size: 128
seed: 42
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
num_epochs: 30
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss
No log	1.9763	500	3.1892
No log	3.9526	1000	2.6822
No log	5.9289	1500	2.4790
3.2335	7.9051	2000	2.3633
3.2335	9.8814	2500	2.2929
3.2335	11.8577	3000	2.2464
3.2335	13.8340	3500	2.2120
2.1358	15.8103	4000	2.1875
2.1358	17.7866	4500	2.1714
2.1358	19.7628	5000	2.1574
2.1358	21.7391	5500	2.1491
1.94	23.7154	6000	2.1411
1.94	25.6917	6500	2.1356
1.94	27.6680	7000	2.1331
1.94	29.6443	7500	2.1313

Khanalnishan/full_nepali-captioning

full_nepali-captioning