ASR-cv-corpus-ug-14

This model is a fine-tuned version of piyazon/ASR-cv-corpus-ug-13 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.0034
• Wer: 0.0033

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: 32
• eval_batch_size: 8
• seed: 42
• 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: 400
• num_epochs: 10
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0033	0.2560	500	0.0061	0.0088
0.0065	0.5120	1000	0.0050	0.0078
0.0056	0.7680	1500	0.0040	0.0062
0.0051	1.0241	2000	0.0056	0.0077
0.0041	1.2801	2500	0.0054	0.0075
0.0035	1.5361	3000	0.0043	0.0065
0.0041	1.7921	3500	0.0051	0.0076
0.0042	2.0481	4000	0.0040	0.0054
0.003	2.3041	4500	0.0040	0.0063
0.0027	2.5602	5000	0.0047	0.0069
0.0037	2.8162	5500	0.0050	0.0073
0.0031	3.0722	6000	0.0038	0.0058
0.0018	3.3282	6500	0.0036	0.0054
0.0022	3.5842	7000	0.0039	0.0053
0.0019	3.8402	7500	0.0043	0.0055
0.0017	4.0963	8000	0.0042	0.0052
0.0021	4.3523	8500	0.0042	0.0056
0.002	4.6083	9000	0.0053	0.0068
0.0027	4.8643	9500	0.0052	0.0071
0.0015	5.1203	10000	0.0038	0.0047
0.001	5.3763	10500	0.0036	0.0042
0.0013	5.6324	11000	0.0035	0.0041
0.001	5.8884	11500	0.0038	0.0051
0.001	6.1444	12000	0.0041	0.0046
0.0011	6.4004	12500	0.0036	0.0043
0.0006	6.6564	13000	0.0036	0.0045
0.0008	6.9124	13500	0.0036	0.0046
0.0005	7.1685	14000	0.0035	0.0037
0.0004	7.4245	14500	0.0035	0.0040
0.0005	7.6805	15000	0.0033	0.0040
0.0006	7.9365	15500	0.0031	0.0036
0.0003	8.1925	16000	0.0033	0.0038
0.0004	8.4485	16500	0.0032	0.0038
0.0003	8.7046	17000	0.0031	0.0036
0.0002	8.9606	17500	0.0030	0.0037
0.0001	9.2166	18000	0.0032	0.0035
0.0001	9.4726	18500	0.0034	0.0034
0.0001	9.7286	19000	0.0034	0.0033
0.0001	9.9846	19500	0.0034	0.0033

Framework versions

• Transformers 4.56.1
• Pytorch 2.8.0+cu128
• Datasets 4.1.0
• Tokenizers 0.22.0