README.md

metadata

license: mit
base_model: microsoft/speecht5_tts
tags:
  - generated_from_trainer
model-index:
  - name: speecht5_tts
    results: []

speecht5_tts

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: 0.5154

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: 2
• eval_batch_size: 2
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 500
• training_steps: 10000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
No log	2.72	250	0.6947
0.7897	5.43	500	0.5278
0.7897	8.15	750	0.5005
0.5429	10.87	1000	0.4842
0.5429	13.59	1250	0.4830
0.4997	16.3	1500	0.4912
0.4997	19.02	1750	0.4820
0.4712	21.74	2000	0.4835
0.4712	24.46	2250	0.4840
0.4621	27.17	2500	0.4925
0.4621	29.89	2750	0.4855
0.4507	32.61	3000	0.4882
0.4507	35.33	3250	0.5065
0.4445	38.04	3500	0.5033
0.4445	40.76	3750	0.4991
0.4345	43.48	4000	0.4950
0.4345	46.2	4250	0.4986
0.4306	48.91	4500	0.5026
0.4306	51.63	4750	0.4986
0.4272	54.35	5000	0.5048
0.4272	57.07	5250	0.4967
0.4234	59.78	5500	0.5011
0.4234	62.5	5750	0.5017
0.4187	65.22	6000	0.5047
0.4187	67.93	6250	0.5041
0.4188	70.65	6500	0.5064
0.4188	73.37	6750	0.5164
0.4108	76.09	7000	0.5133
0.4108	78.8	7250	0.5086
0.4118	81.52	7500	0.5070
0.4118	84.24	7750	0.5093
0.4082	86.96	8000	0.5155
0.4082	89.67	8250	0.5089
0.407	92.39	8500	0.5134
0.407	95.11	8750	0.5056
0.407	97.83	9000	0.5154
0.407	100.54	9250	0.5108
0.4062	103.26	9500	0.5112
0.4062	105.98	9750	0.5122
0.4052	108.7	10000	0.5154

Framework versions

• Transformers 4.36.0.dev0
• Pytorch 2.1.0+cu121
• Datasets 2.15.0
• Tokenizers 0.14.1