zlm_b64_le4_s8000

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: 0.3177

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: 42
• gradient_accumulation_steps: 8
• 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: 8000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.5277	0.4188	500	0.4806
0.4582	0.8377	1000	0.4116
0.4312	1.2565	1500	0.3951
0.4122	1.6754	2000	0.3768
0.4002	2.0942	2500	0.3599
0.3905	2.5131	3000	0.3521
0.3806	2.9319	3500	0.3445
0.37	3.3508	4000	0.3474
0.3736	3.7696	4500	0.3362
0.3608	8.3872	5000	0.3342
0.3602	9.2249	5500	0.3258
0.3561	10.0626	6000	0.3230
0.3505	10.9003	6500	0.3199
0.3473	11.7380	7000	0.3193
0.3523	12.5757	7500	0.3177
0.3462	13.4134	8000	0.3177

Framework versions

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