wav2vec2-lv-60-espeak-cv-ft-WCTCv2-phocab-ds-f9

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: 20.3722
• Per: 0.0514

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
1318.5509	0.7194	400	274.4972	1.0
518.9905	1.4388	800	264.3368	1.0
498.0993	2.1583	1200	245.4142	1.0
386.328	2.8777	1600	100.2428	0.3954
213.4631	3.5971	2000	54.9563	0.1948
159.9631	4.3165	2400	40.6919	0.1214
135.1337	5.0360	2800	36.6996	0.0911
122.5109	5.7554	3200	32.3255	0.0801
112.5183	6.4748	3600	29.0416	0.0742
108.7395	7.1942	4000	27.0754	0.0691
103.0256	7.9137	4400	25.5572	0.0708
97.8429	8.6331	4800	24.8682	0.0649
95.4275	9.3525	5200	23.1266	0.0666
93.7924	10.0719	5600	23.5796	0.0691
90.5355	10.7914	6000	21.9151	0.0632
87.9662	11.5108	6400	22.9246	0.0658
88.6945	12.2302	6800	21.4374	0.0616
84.5592	12.9496	7200	21.0339	0.0624
84.8458	13.6691	7600	22.6146	0.0649
80.5781	14.3885	8000	20.6335	0.0641
81.6724	15.1079	8400	20.4397	0.0607
80.869	15.8273	8800	20.3979	0.0624
82.8611	16.5468	9200	22.4343	0.0599
79.3414	17.2662	9600	23.2001	0.0616
79.5972	17.9856	10000	20.4888	0.0573
77.7815	18.7050	10400	21.2219	0.0590
78.0923	19.4245	10800	23.4833	0.0649
78.0261	20.1439	11200	22.5931	0.0590
74.4771	20.8633	11600	20.9137	0.0540
76.4859	21.5827	12000	20.6029	0.0590
74.8805	22.3022	12400	21.0946	0.0531
73.7169	23.0216	12800	20.6823	0.0556
74.8553	23.7410	13200	19.8081	0.0548
74.2606	24.4604	13600	20.3502	0.0531
72.0619	25.1799	14000	19.7951	0.0506
71.35	25.8993	14400	20.3513	0.0514
70.7273	26.6187	14800	21.6780	0.0523
72.5879	27.3381	15200	20.5299	0.0506
71.3647	28.0576	15600	18.9935	0.0514
69.9942	28.7770	16000	19.4889	0.0506
72.9756	29.4964	16400	20.3722	0.0514

Framework versions

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