wav2vec2-lv-60-espeak-cv-ft-WCTC-phocab-ds-f8

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: 11.8064
• Per: 0.0451

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
1388.9211	0.7194	400	258.8207	1.0
496.2495	1.4388	800	241.3190	1.0
467.8013	2.1583	1200	209.6756	0.9895
319.2299	2.8777	1600	64.0295	0.2880
171.572	3.5971	2000	38.2525	0.1529
133.5128	4.3165	2400	29.6704	0.1030
114.4326	5.0360	2800	24.8591	0.0853
101.9263	5.7554	3200	21.7826	0.0764
94.2615	6.4748	3600	19.1824	0.0692
90.1317	7.1942	4000	18.4454	0.0708
84.2694	7.9137	4400	17.8917	0.0668
82.0402	8.6331	4800	16.9460	0.0652
79.0095	9.3525	5200	16.8115	0.0652
77.0721	10.0719	5600	16.3140	0.0619
73.6757	10.7914	6000	15.6164	0.0636
70.9067	11.5108	6400	15.7822	0.0611
72.3999	12.2302	6800	14.8963	0.0603
70.6939	12.9496	7200	14.0957	0.0571
69.9356	13.6691	7600	13.9708	0.0523
67.5808	14.3885	8000	13.7943	0.0523
68.9874	15.1079	8400	13.7925	0.0523
63.5566	15.8273	8800	13.4886	0.0531
64.1214	16.5468	9200	13.1847	0.0523
64.8615	17.2662	9600	13.5771	0.0547
62.7331	17.9856	10000	13.7224	0.0531
61.7567	18.7050	10400	13.6806	0.0531
62.0574	19.4245	10800	12.6783	0.0515
59.3861	20.1439	11200	13.1509	0.0507
61.258	20.8633	11600	12.6506	0.0483
58.7836	21.5827	12000	12.7024	0.0491
59.459	22.3022	12400	12.1852	0.0491
58.4748	23.0216	12800	12.5531	0.0475
59.3281	23.7410	13200	12.1136	0.0442
59.496	24.4604	13600	12.0433	0.0467
57.9776	25.1799	14000	12.1108	0.0483
56.2327	25.8993	14400	11.9060	0.0467
58.9476	26.6187	14800	11.8183	0.0467
57.7004	27.3381	15200	11.9450	0.0451
57.3872	28.0576	15600	11.8708	0.0451
54.7268	28.7770	16000	11.9073	0.0467
57.4947	29.4964	16400	11.8064	0.0451

Framework versions

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