w2v-bert-2.0-ml-superb-xty

This model is a fine-tuned version of facebook/w2v-bert-2.0 on the ml-superb-subset dataset. It achieves the following results on the evaluation set:

• Loss: 2.3981
• Wer: 1.3985

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: 30
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
3.5467	0.8219	30	2.8636	1.0
2.4639	1.6438	60	2.5298	1.0094
2.38	2.4658	90	2.4983	1.1263
2.2725	3.2877	120	2.4866	1.2319
2.2608	4.1096	150	2.5116	1.5405
2.2222	4.9315	180	2.4588	1.3300
2.2609	5.7534	210	2.4448	1.3451
2.1665	6.5753	240	2.4270	1.3199
2.1703	7.3973	270	2.4223	1.3576
2.1366	8.2192	300	2.4054	1.4085
2.123	9.0411	330	2.4006	1.4180
2.1331	9.8630	360	2.3981	1.3985

Framework versions

• Transformers 4.40.2
• Pytorch 2.2.1+cu121
• Datasets 2.19.1
• Tokenizers 0.19.1