learning_rate: 5e-05
train_batch_size: 16
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 32
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
lr_scheduler_warmup_steps: 500
num_epochs: 10
mixed_precision_training: Native AMP

Training Loss	Epoch	Step	Validation Loss	Cer
5.5363	0.6192	300	1.0539	0.2298
1.8453	1.2374	600	0.9015	0.2054
1.4572	1.8566	900	0.8487	0.1915
1.1530	2.4747	1200	0.8267	0.1961
0.9653	3.0929	1500	0.8389	0.1831
0.7378	3.7121	1800	0.8667	0.1791
0.5700	4.3302	2100	0.9389	0.1802
0.4421	4.9494	2400	1.0079	0.1863
0.2986	5.5676	2700	1.1053	0.1783
0.2296	6.1858	3000	1.3608	0.1759
0.1616	6.8050	3300	1.3994	0.1765
0.1144	7.4231	3600	1.5226	0.1768
0.0825	8.0413	3900	1.6242	0.1762
0.0498	8.6605	4200	1.7148	0.1762
0.0385	9.2786	4500	1.8250	0.1748
0.0247	9.8978	4800	1.8841	0.1753

dhasmana/Chhattisgarhi-Magadhi-Marathi-Rajasthani-w2v-bert-2.0

Chhattisgarhi-Magadhi-Marathi-Rajasthani-w2v-bert-2.0