results

This model is a fine-tuned version of ai4bharat/IndicBART on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0863
• Wer: 0.4892

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 2e-05
• train_batch_size: 64
• eval_batch_size: 64
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• num_epochs: 4

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
1.4597	0.4504	4000	0.3822	0.6293
0.3574	0.9008	8000	0.2432	0.5683
0.252	1.3512	12000	0.1656	0.5286
0.1921	1.8016	16000	0.1277	0.5102
0.1583	2.2520	20000	0.1074	0.5010
0.1383	2.7024	24000	0.0967	0.4937
0.1266	3.1528	28000	0.0896	0.4904
0.1199	3.6032	32000	0.0863	0.4892

Framework versions

• Transformers 4.45.2
• Pytorch 2.1.1+cu121
• Datasets 3.0.1
• Tokenizers 0.20.1