senga-LUK-20k-nodegate-speecht5

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: 0.2060

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: 2000
• training_steps: 20000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.2064	90.9091	1000	0.2125
0.1512	181.8182	2000	0.1874
0.1226	272.7273	3000	0.1677
0.1094	363.6364	4000	0.1826
0.0985	454.5455	5000	0.1938
0.0944	545.4545	6000	0.1773
0.0861	636.3636	7000	0.1857
0.0777	727.2727	8000	0.1889
0.0832	818.1818	9000	0.2049
0.0738	909.0909	10000	0.1930
0.0651	1000.0	11000	0.1955
0.0679	1090.9091	12000	0.1939
0.0619	1181.8182	13000	0.1964
0.0632	1272.7273	14000	0.2025
0.0568	1363.6364	15000	0.2065
0.0587	1454.5455	16000	0.2050
0.056	1545.4545	17000	0.2086
0.0709	1636.3636	18000	0.2143
0.0573	1727.2727	19000	0.2085
0.0589	1818.1818	20000	0.2060

Framework versions

• Transformers 4.57.1
• Pytorch 2.8.0+cu128
• Datasets 4.2.0
• Tokenizers 0.22.2