wav2vec2-lv-60-espeak-cv-ft-WCTCv2-phocab-ds-f7

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: 14.2388
• Per: 0.0447

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
1438.2527	0.7194	400	262.2246	1.0
521.9669	1.4388	800	250.0486	1.0
492.9216	2.1583	1200	215.9474	0.9908
415.6159	2.8777	1600	155.8855	0.7912
294.7445	3.5971	2000	76.1166	0.3328
196.8822	4.3165	2400	39.4806	0.1456
150.387	5.0360	2800	31.9243	0.1002
131.5259	5.7554	3200	26.9720	0.0655
119.266	6.4748	3600	24.0607	0.0624
111.1528	7.1942	4000	24.3722	0.0532
107.8178	7.9137	4400	21.8196	0.0478
99.3233	8.6331	4800	21.1698	0.0601
100.8546	9.3525	5200	20.1011	0.0516
94.4388	10.0719	5600	20.4904	0.0562
95.4749	10.7914	6000	19.3444	0.0501
89.2203	11.5108	6400	18.9132	0.0501
88.5827	12.2302	6800	18.0343	0.0447
87.5782	12.9496	7200	17.4712	0.0493
89.7401	13.6691	7600	17.4045	0.0508
82.1285	14.3885	8000	17.5382	0.0478
84.9758	15.1079	8400	16.7021	0.0493
81.8318	15.8273	8800	16.4826	0.0485
82.9694	16.5468	9200	15.7187	0.0478
79.7547	17.2662	9600	16.0707	0.0470
80.564	17.9856	10000	16.1449	0.0455
78.5287	18.7050	10400	15.5053	0.0470
78.3303	19.4245	10800	15.2458	0.0462
77.5218	20.1439	11200	15.8553	0.0470
78.8467	20.8633	11600	14.4955	0.0393
76.9731	21.5827	12000	15.0185	0.0439
76.7753	22.3022	12400	15.6488	0.0431
75.1212	23.0216	12800	15.2574	0.0455
75.2347	23.7410	13200	15.2690	0.0462
74.1773	24.4604	13600	15.0007	0.0447
72.2866	25.1799	14000	14.5996	0.0424
74.1712	25.8993	14400	14.5717	0.0455
71.8524	26.6187	14800	14.2555	0.0462
75.3872	27.3381	15200	13.8649	0.0408
71.5674	28.0576	15600	13.9912	0.0424
71.6251	28.7770	16000	14.0903	0.0455
72.3035	29.4964	16400	14.2388	0.0447

Framework versions

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