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: 0.6815

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: 4
eval_batch_size: 4
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: 30000
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0.53	250	1.1437
1.3289	1.06	500	0.8521
1.3289	1.6	750	0.7901
0.8977	2.13	1000	0.7478
0.8977	2.66	1250	0.7437
0.8131	3.19	1500	0.7243
0.8131	3.72	1750	0.7106
0.771	4.26	2000	0.7072
0.771	4.79	2250	0.7008
0.7562	5.32	2500	0.6916
0.7562	5.85	2750	0.6850
0.7472	6.38	3000	0.6876
0.7472	6.91	3250	0.6807
0.7266	7.45	3500	0.6804
0.7266	7.98	3750	0.6763
0.715	8.51	4000	0.6769
0.715	9.04	4250	0.6698
0.7005	9.57	4500	0.6690
0.7005	10.11	4750	0.6653
0.6932	10.64	5000	0.6656
0.6932	11.17	5250	0.6684
0.6854	11.7	5500	0.6645
0.6854	12.23	5750	0.6634
0.6739	12.77	6000	0.6674
0.6739	13.3	6250	0.6606
0.6754	13.83	6500	0.6663
0.6754	14.36	6750	0.6681
0.6592	14.89	7000	0.6589
0.6592	15.43	7250	0.6601
0.6528	15.96	7500	0.6739
0.6528	16.49	7750	0.6643
0.6539	17.02	8000	0.6605
0.6539	17.55	8250	0.6614
0.6437	18.09	8500	0.6551
0.6437	18.62	8750	0.6604
0.6341	19.15	9000	0.6606
0.6341	19.68	9250	0.6582
0.6305	20.21	9500	0.6714
0.6305	20.74	9750	0.6618
0.627	21.28	10000	0.6600
0.627	21.81	10250	0.6636
0.6244	22.34	10500	0.6692
0.6244	22.87	10750	0.6645
0.6178	23.4	11000	0.6670
0.6178	23.94	11250	0.6611
0.6157	24.47	11500	0.6697
0.6157	25.0	11750	0.6651
0.6108	25.53	12000	0.6642
0.6108	26.06	12250	0.6646
0.6008	26.6	12500	0.6672
0.6008	27.13	12750	0.6601
0.6067	27.66	13000	0.6760
0.6067	28.19	13250	0.6639
0.5985	28.72	13500	0.6662
0.5985	29.26	13750	0.6720
0.5957	29.79	14000	0.6710
0.5957	30.32	14250	0.6688
0.5944	30.85	14500	0.6714
0.5944	31.38	14750	0.6760
0.5886	31.91	15000	0.6639
0.5886	32.45	15250	0.6714
0.5868	32.98	15500	0.6722
0.5868	33.51	15750	0.6790
0.5851	34.04	16000	0.6728
0.5851	34.57	16250	0.6812
0.5819	35.11	16500	0.6756
0.5819	35.64	16750	0.6679
0.5811	36.17	17000	0.6719
0.5811	36.7	17250	0.6684
0.5759	37.23	17500	0.6776
0.5759	37.77	17750	0.6743
0.5743	38.3	18000	0.6725
0.5743	38.83	18250	0.6730
0.5761	39.36	18500	0.6712
0.5761	39.89	18750	0.6765
0.576	40.43	19000	0.6779
0.576	40.96	19250	0.6801
0.5734	41.49	19500	0.6756
0.5734	42.02	19750	0.6761
0.5743	42.55	20000	0.6857
0.5743	43.09	20250	0.6734
0.5732	43.62	20500	0.6753
0.5732	44.15	20750	0.6803
0.5657	44.68	21000	0.6743
0.5657	45.21	21250	0.6831
0.565	45.74	21500	0.6799
0.565	46.28	21750	0.6769
0.565	46.81	22000	0.6786
0.565	47.34	22250	0.6788
0.5583	47.87	22500	0.6830
0.5583	48.4	22750	0.6884
0.5652	48.94	23000	0.6827
0.5652	49.47	23250	0.6795
0.5625	50.0	23500	0.6807
0.5625	50.53	23750	0.6788
0.5605	51.06	24000	0.6862
0.5605	51.6	24250	0.6822
0.5571	52.13	24500	0.6819
0.5571	52.66	24750	0.6797
0.5633	53.19	25000	0.6835
0.5633	53.72	25250	0.6835
0.5572	54.26	25500	0.6881
0.5572	54.79	25750	0.6791
0.5571	55.32	26000	0.6815
0.5571	55.85	26250	0.6868
0.5534	56.38	26500	0.6876
0.5534	56.91	26750	0.6871
0.5525	57.45	27000	0.6836
0.5525	57.98	27250	0.6841
0.5542	58.51	27500	0.6911
0.5542	59.04	27750	0.6835
0.5512	59.57	28000	0.6806
0.5512	60.11	28250	0.6805
0.5474	60.64	28500	0.6858
0.5474	61.17	28750	0.6874
0.5548	61.7	29000	0.6811
0.5548	62.23	29250	0.6808
0.5545	62.77	29500	0.6868
0.5545	63.3	29750	0.6894
0.5522	63.83	30000	0.6815

Framework versions

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