ceb_b64_le5_s12000

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: 0.3942

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: 8
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 2000
• training_steps: 12000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.5368	19.8020	500	0.4637
0.4674	39.6040	1000	0.4246
0.448	59.4059	1500	0.4101
0.4358	79.2079	2000	0.4011
0.4273	99.0099	2500	0.3966
0.4161	118.8119	3000	0.3953
0.4119	138.6139	3500	0.3939
0.4008	158.4158	4000	0.3932
0.4029	178.2178	4500	0.3932
0.4008	198.0198	5000	0.3947
0.3989	217.8218	5500	0.3933
0.3934	237.6238	6000	0.3943
0.3927	257.4257	6500	0.3937
0.3875	277.2277	7000	0.3918
0.3914	297.0297	7500	0.3931
0.3911	316.8317	8000	0.3937
0.3821	336.6337	8500	0.3925
0.3932	356.4356	9000	0.3938
0.3861	376.2376	9500	0.3940
0.3829	396.0396	10000	0.3934
0.3819	415.8416	10500	0.3935
0.383	435.6436	11000	0.3939
0.3828	455.4455	11500	0.3937
0.3818	475.2475	12000	0.3942

Framework versions

• Transformers 4.41.0.dev0
• Pytorch 2.2.1+cu121
• Datasets 2.19.0
• Tokenizers 0.19.1