STT_Model_4

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

• Loss: 0.2311
• Wer: 0.1373

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: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 1000
• num_epochs: 100

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
3.4196	5.68	500	0.9866	0.6983
0.3696	11.36	1000	0.8788	0.4010
0.1182	17.05	1500	0.2187	0.1947
0.0658	22.73	2000	0.2578	0.1757
0.0421	28.41	2500	0.2178	0.1609
0.0346	34.09	3000	0.2038	0.1584
0.0285	39.77	3500	0.2187	0.1594
0.0228	45.45	4000	0.2114	0.1445
0.0262	51.14	4500	0.2201	0.1631
0.0162	56.82	5000	0.2078	0.1424
0.0135	62.5	5500	0.1989	0.1393
0.0128	68.18	6000	0.2118	0.1410
0.0104	73.86	6500	0.2158	0.1361
0.0081	79.55	7000	0.2154	0.1348
0.0067	85.23	7500	0.2107	0.1358
0.0067	90.91	8000	0.2161	0.1373
0.0056	96.59	8500	0.2311	0.1373

Framework versions

• Transformers 4.26.0
• Pytorch 1.13.1+cu116
• Datasets 2.9.0
• Tokenizers 0.13.2