moonshine_tiny_pt_v02

This model is a fine-tuned version of aomocelin/moonshine_tiny_pt on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 11.6990
• Wer: 33.4344

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: 1e-05
• train_batch_size: 16
• eval_batch_size: 64
• seed: 42
• optimizer: Use OptimizerNames.ADAMW_TORCH 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: 0.03
• training_steps: 15000
• mixed_precision_training: Native AMP
• label_smoothing_factor: 0.1

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
2.1035	0.0469	500	11.6412	36.6414
2.1216	0.0938	1000	11.5398	35.9324
2.0764	0.1407	1500	11.5108	35.5104
2.0585	0.1876	2000	11.5237	35.5429
2.0636	0.2345	2500	11.5384	35.4225
2.0700	0.2814	3000	11.4961	35.1111
2.0735	0.3283	3500	11.5295	34.8330
2.0478	0.3752	4000	11.6607	34.6289
2.0577	0.4221	4500	11.5908	34.6582
2.0391	0.4690	5000	11.5946	34.3947
2.0299	0.5159	5500	11.6431	34.5785
2.0199	0.5629	6000	11.6115	34.1947
2.0414	0.6098	6500	11.6344	34.4435
2.0228	0.6567	7000	11.6368	34.1898
2.0414	0.7036	7500	11.6024	33.8508
2.0231	0.7505	8000	11.6136	33.8760
2.0074	0.7974	8500	11.6744	33.8963
2.0224	0.8443	9000	11.6394	33.5223
2.0123	0.8912	9500	11.6156	33.5881
2.0303	0.9381	10000	11.6476	33.7597
2.0475	0.9850	10500	11.6160	33.6483
1.9851	1.0319	11000	11.6783	33.5686
1.9642	1.0788	11500	11.6799	33.6645
2.0111	1.1257	12000	11.6853	33.8752
1.9904	1.1726	12500	11.7049	33.3564
1.9923	1.2195	13000	11.6766	33.5166
2.0205	1.2664	13500	11.7066	33.4052
2.0005	1.3133	14000	11.6980	33.5499
1.9924	1.3602	14500	11.7034	33.4767
1.9653	1.4071	15000	11.6990	33.4344

Framework versions

• Transformers 5.12.1
• Pytorch 2.11.0+cu128
• Datasets 5.0.0
• Tokenizers 0.22.2