speecht5_finetuned_massiveTR

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: 0.4373

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: 4
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 8
• total_train_batch_size: 32
• 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: 100
• training_steps: 2300
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
4.08	28.9231	115	0.6232
3.0688	57.6154	230	0.4632
2.8275	86.3077	345	0.4434
2.6628	115.0	460	0.4378
2.6075	143.9231	575	0.4441
2.5445	172.6154	690	0.4342
2.4239	201.3077	805	0.4325
2.3723	230.0	920	0.4553
2.4305	258.9231	1035	0.4384
2.4016	287.6154	1150	0.4322
2.4025	316.3077	1265	0.4454
2.329	345.0	1380	0.4438
2.3223	373.9231	1495	0.4398
2.2315	402.6154	1610	0.4415
2.3544	431.3077	1725	0.4334
2.2827	460.0	1840	0.4381
2.3045	488.9231	1955	0.4482
2.2728	517.6154	2070	0.4575
2.2962	546.3077	2185	0.4516
2.2531	575.0	2300	0.4373

Framework versions

• Transformers 4.47.1
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0