w2v-bert-Marathi-large

This model is a fine-tuned version of facebook/w2v-bert-2.0 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.2714
• Wer: 0.1698
• Cer: 0.0531

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: 5e-05
• train_batch_size: 2
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 8
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 3000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
2.8852	0.5882	300	0.7826	0.4911	0.1647
0.6243	1.1765	600	0.6280	0.3920	0.1351
0.4901	1.7647	900	0.4369	0.3101	0.0986
0.355	2.3529	1200	0.3922	0.2658	0.0849
0.2943	2.9412	1500	0.3400	0.2371	0.0753
0.2177	3.5294	1800	0.3041	0.2080	0.0646
0.1779	4.1176	2100	0.2906	0.1954	0.0608
0.1299	4.7059	2400	0.2904	0.1779	0.0560
0.0929	5.2941	2700	0.2885	0.1727	0.0537
0.0729	5.8824	3000	0.2714	0.1698	0.0531

Framework versions

• Transformers 4.42.0.dev0
• Pytorch 2.3.0+cu121
• Datasets 2.19.2
• Tokenizers 0.19.1