myx_tts_lsm

This model is a fine-tuned version of microsoft/speecht5_tts on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0551

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: 0.0001
• train_batch_size: 8
• eval_batch_size: 8
• seed: 3407
• gradient_accumulation_steps: 4
• 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: cosine
• lr_scheduler_warmup_steps: 1500
• training_steps: 15000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.0943	3.8176	1000	0.0708
0.0872	7.6342	2000	0.0700
0.0802	11.4508	3000	0.0613
0.0777	15.2674	4000	0.0613
0.0748	19.0840	5000	0.0596
0.0725	22.9016	6000	0.0588
0.0706	26.7182	7000	0.0575
0.0697	30.5349	8000	0.0579
0.0671	34.3515	9000	0.0564
0.065	38.1681	10000	0.0559
0.0636	41.9857	11000	0.0556
0.0617	45.8023	12000	0.0554
0.0622	49.6189	13000	0.0554
0.0635	53.4355	14000	0.0554
0.0624	57.2521	15000	0.0551

Framework versions

• Transformers 4.57.1
• Pytorch 2.8.0+cu128
• Datasets 3.6.0
• Tokenizers 0.22.2