Modelpublic

Birma/speecht5_finetuned_zarma

sourceHugging Facemitupdated 9mo agoView on Hugging Face

0likes8downloads

Model Card

speecht5finetunedzarma

This model is a fine-tuned version of microsoft/speecht5_tts on an unknown dataset. It achieves the following results on the evaluation set:

—Loss: nan

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
—trainbatchsize: 4
—evalbatchsize: 2
—seed: 42
—gradientaccumulationsteps: 8
—totaltrainbatch_size: 32
—optimizer: Use OptimizerNames.ADAMWTORCHFUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
—lrschedulertype: linear
—lrschedulerwarmup_steps: 50
—training_steps: 1000
—mixedprecisiontraining: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
2.2591	16.8889	50	nan
1.7534	33.4444	100	nan
1.446	50.0	150	nan
1.3861	66.8889	200	nan
1.2551	83.4444	250	nan
1.1811	100.0	300	nan
1.1416	116.8889	350	nan
1.0839	133.4444	400	nan
1.0502	150.0	450	nan
1.0335	166.8889	500	nan
1.0372	183.4444	550	nan
1.0037	200.0	600	nan
0.9739	216.8889	650	nan
1.027	233.4444	700	nan
0.9747	250.0	750	nan
0.9857	266.8889	800	nan
0.9798	283.4444	850	nan
0.9619	300.0	900	nan
0.9532	316.8889	950	nan
0.9474	333.4444	1000	nan

Framework versions

—Transformers 4.57.3
—Pytorch 2.9.0+cu126
—Datasets 4.0.0
—Tokenizers 0.22.1