macarious/torgo_xlsr_finetune_F03_old

torgo_xlsr_finetune_F03

This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.3758
• Wer: 0.1234

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: 0.0001
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• gradient_accumulation_steps: 2
• 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: 1000
• num_epochs: 20

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
3.5793	1.06	1000	3.5545	1.0
2.1557	2.11	2000	0.9675	0.7402
0.9155	3.17	3000	0.5581	0.4278
0.6488	4.23	4000	0.4753	0.2703
0.4806	5.29	5000	0.5099	0.1864
0.4012	6.34	6000	0.3734	0.1549
0.3717	7.4	7000	0.4955	0.1732
0.3247	8.46	8000	0.4187	0.1339
0.3028	9.51	9000	0.3920	0.1234
0.2716	10.57	10000	0.6216	0.1785
0.255	11.63	11000	0.3310	0.1417
0.2418	12.68	12000	0.3653	0.1339
0.2111	13.74	13000	0.3677	0.1155
0.1996	14.8	14000	0.4039	0.1286
0.1869	15.86	15000	0.3811	0.1339
0.175	16.91	16000	0.3736	0.1286
0.1651	17.97	17000	0.3848	0.1260
0.179	19.03	18000	0.3758	0.1234

Framework versions

• Transformers 4.26.1
• Pytorch 2.2.0
• Datasets 2.16.1
• Tokenizers 0.13.3