wav2vec2-lv-60-espeak-cv-ft-WCTCv2-phocab-ds-f6

This model is a fine-tuned version of facebook/wav2vec2-lv-60-espeak-cv-ft on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 17.6479
• Per: 0.0423

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: 3e-05
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• num_epochs: 30
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Per
1444.1388	0.7194	400	273.8882	1.0
528.9609	1.4388	800	256.7122	1.0
501.5803	2.1583	1200	242.6743	1.0
417.1812	2.8777	1600	117.0687	0.5390
222.6512	3.5971	2000	49.3962	0.1675
166.2119	4.3165	2400	36.1953	0.1299
139.3321	5.0360	2800	29.4526	0.0934
127.0769	5.7554	3200	26.2651	0.0799
116.3417	6.4748	3600	23.4358	0.0693
112.0814	7.1942	4000	21.1188	0.0693
106.5399	7.9137	4400	20.4718	0.0626
101.5809	8.6331	4800	19.3273	0.0587
101.282	9.3525	5200	18.0581	0.0568
94.5719	10.0719	5600	17.2480	0.0529
92.766	10.7914	6000	17.1745	0.0587
87.9467	11.5108	6400	18.0644	0.0481
89.8651	12.2302	6800	17.9260	0.0520
89.2698	12.9496	7200	17.4550	0.0520
86.7451	13.6691	7600	17.4322	0.0481
86.0477	14.3885	8000	17.2115	0.0529
86.5714	15.1079	8400	17.5130	0.0510
83.7452	15.8273	8800	18.6866	0.0539
83.7124	16.5468	9200	18.3658	0.0520
81.1891	17.2662	9600	17.4496	0.0491
82.0165	17.9856	10000	17.9179	0.0452
79.8557	18.7050	10400	18.1328	0.0491
79.4204	19.4245	10800	17.8189	0.0462
79.172	20.1439	11200	17.1571	0.0500
79.8598	20.8633	11600	17.2944	0.0423
77.7783	21.5827	12000	16.9868	0.0462
75.899	22.3022	12400	17.7540	0.0423
77.5824	23.0216	12800	17.8216	0.0443
75.8729	23.7410	13200	17.8142	0.0414
77.4199	24.4604	13600	17.8511	0.0423
75.0168	25.1799	14000	17.4843	0.0404
75.5704	25.8993	14400	17.7559	0.0395
74.903	26.6187	14800	18.2069	0.0404
74.6692	27.3381	15200	18.0912	0.0414
76.1732	28.0576	15600	18.5670	0.0414
73.5948	28.7770	16000	17.8981	0.0423
72.1325	29.4964	16400	17.6479	0.0423

Framework versions

• Transformers 4.57.6
• Pytorch 2.9.1+cu128
• Datasets 4.5.0
• Tokenizers 0.22.2