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
6.3624	0.5420	300	1.2186	0.2870
2.0074	1.0831	600	1.0891	0.2836
1.5654	1.6251	900	1.0085	0.2392
1.2836	2.1662	1200	1.0369	0.2312
1.0187	2.7082	1500	1.0477	0.2263
0.7801	3.2493	1800	1.1672	0.2243
0.6002	3.7913	2100	1.1441	0.2357
0.4305	4.3324	2400	1.3504	0.2292
0.3316	4.8744	2700	1.4374	0.2295
0.2195	5.4155	3000	1.5483	0.2266
0.1747	5.9575	3300	1.6746	0.2252
0.1101	6.4986	3600	1.8221	0.2272
0.0897	7.0397	3900	1.9780	0.2249
0.0566	7.5818	4200	2.0282	0.2252
0.0408	8.1229	4500	2.2570	0.2230
0.0247	8.6649	4800	2.2963	0.2242
0.0168	9.2060	5100	2.4469	0.2219
0.0102	9.7480	5400	2.5161	0.2210

dhasmana/Chhattisgarhi-Magadhi-Marwari-Garhwali-Rajasthani-w2v-bert-2.0

Chhattisgarhi-Magadhi-Marwari-Garhwali-Rajasthani-w2v-bert-2.0