intit_model

This model is a fine-tuned version of facebook/wav2vec2-base-960h on the None dataset. It achieves the following results on the evaluation set:

• Loss: 2.2486
• Wer: 0.4348
• Cer: 0.9047

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: 1e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• training_steps: 1000

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
0.9753	20.0	100	1.3804	0.5072	0.9054
0.5395	40.0	200	1.5495	0.4444	0.9062
0.3735	60.0	300	1.7729	0.4396	0.9056
0.2427	80.0	400	1.9016	0.4348	0.9063
0.2389	100.0	500	2.0569	0.4348	0.9061
0.1822	120.0	600	2.0684	0.4300	0.9050
0.1578	140.0	700	2.1332	0.4396	0.9049
0.1547	160.0	800	2.2138	0.4444	0.9047
0.1807	180.0	900	2.2467	0.4348	0.9047
0.1427	200.0	1000	2.2486	0.4348	0.9047

Framework versions

• Transformers 4.35.2
• Pytorch 2.1.0+cu118
• Datasets 2.15.0
• Tokenizers 0.15.0