wav2vec2-lv-60-espeak-cv-ft-WCTCv2-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: 18.3136
• Per: 0.0483

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
1426.2453	0.7194	400	277.0271	1.0
524.2816	1.4388	800	260.1468	1.0
498.9254	2.1583	1200	240.6920	1.0
375.756	2.8777	1600	105.9157	0.4803
212.6498	3.5971	2000	52.0211	0.1625
156.0128	4.3165	2400	39.2742	0.1094
132.4803	5.0360	2800	34.1151	0.0780
119.1087	5.7554	3200	31.0236	0.0740
111.2812	6.4748	3600	27.9523	0.0660
106.4233	7.1942	4000	26.1573	0.0668
100.2342	7.9137	4400	26.4386	0.0652
97.6576	8.6331	4800	24.4673	0.0628
95.6419	9.3525	5200	24.5329	0.0660
91.3951	10.0719	5600	23.0837	0.0563
88.9433	10.7914	6000	22.6761	0.0619
87.3779	11.5108	6400	22.5947	0.0547
88.0013	12.2302	6800	23.1984	0.0563
86.8282	12.9496	7200	22.3060	0.0539
86.4227	13.6691	7600	22.4437	0.0563
82.7239	14.3885	8000	21.9167	0.0587
85.1721	15.1079	8400	21.3077	0.0563
78.4237	15.8273	8800	21.2354	0.0595
79.8673	16.5468	9200	20.4697	0.0595
80.7659	17.2662	9600	22.2607	0.0571
78.0479	17.9856	10000	20.6864	0.0563
77.2283	18.7050	10400	20.7022	0.0531
77.9888	19.4245	10800	20.5079	0.0523
74.7791	20.1439	11200	19.7176	0.0491
78.0211	20.8633	11600	20.0579	0.0515
73.7585	21.5827	12000	19.9170	0.0547
75.241	22.3022	12400	18.7952	0.0531
73.026	23.0216	12800	18.4493	0.0531
74.0824	23.7410	13200	18.4118	0.0499
74.6452	24.4604	13600	18.0993	0.0475
70.9868	25.1799	14000	18.6080	0.0507
71.1455	25.8993	14400	18.5031	0.0499
73.7071	26.6187	14800	18.4456	0.0499
71.5456	27.3381	15200	18.2258	0.0467
71.9122	28.0576	15600	18.3592	0.0483
69.4866	28.7770	16000	18.2676	0.0483
72.3204	29.4964	16400	18.3136	0.0483

Framework versions

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