speecht5_finetuned_speaking_style_en_2

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: 0.3193

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: 8
• eval_batch_size: 2
• 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: 10000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
0.4074	0.61	100	0.3635
0.4029	1.23	200	0.3629
0.4041	1.84	300	0.3617
0.4006	2.45	400	0.3605
0.3987	3.07	500	0.3563
0.3983	3.68	600	0.3557
0.3949	4.29	700	0.3529
0.388	4.9	800	0.3515
0.3842	5.52	900	0.3484
0.3833	6.13	1000	0.3484
0.3789	6.74	1100	0.3468
0.378	7.36	1200	0.3431
0.3737	7.97	1300	0.3432
0.3737	8.58	1400	0.3432
0.3722	9.2	1500	0.3429
0.3702	9.81	1600	0.3391
0.3672	10.42	1700	0.3373
0.3657	11.03	1800	0.3376
0.3612	11.65	1900	0.3377
0.3615	12.26	2000	0.3327
0.3597	12.87	2100	0.3326
0.358	13.49	2200	0.3317
0.3542	14.1	2300	0.3348
0.3559	14.71	2400	0.3310
0.3567	15.33	2500	0.3335
0.3541	15.94	2600	0.3333
0.3524	16.55	2700	0.3298
0.3494	17.16	2800	0.3287
0.3508	17.78	2900	0.3260
0.3487	18.39	3000	0.3274
0.3484	19.0	3100	0.3295
0.3472	19.62	3200	0.3263
0.3469	20.23	3300	0.3263
0.3454	20.84	3400	0.3280
0.3431	21.46	3500	0.3286
0.3444	22.07	3600	0.3275
0.3435	22.68	3700	0.3281
0.345	23.3	3800	0.3247
0.3438	23.91	3900	0.3263
0.3404	24.52	4000	0.3256
0.342	25.13	4100	0.3273
0.3419	25.75	4200	0.3226
0.34	26.36	4300	0.3218
0.3404	26.97	4400	0.3266
0.3401	27.59	4500	0.3222
0.3398	28.2	4600	0.3236
0.3393	28.81	4700	0.3237
0.3377	29.43	4800	0.3225
0.3374	30.04	4900	0.3236
0.3376	30.65	5000	0.3216
0.3352	31.26	5100	0.3230
0.3367	31.88	5200	0.3208
0.3368	32.49	5300	0.3247
0.3367	33.1	5400	0.3226
0.3375	33.72	5500	0.3203
0.3365	34.33	5600	0.3209
0.3353	34.94	5700	0.3231
0.3352	35.56	5800	0.3201
0.3335	36.17	5900	0.3209
0.334	36.78	6000	0.3204
0.3342	37.39	6100	0.3203
0.3327	38.01	6200	0.3195
0.3342	38.62	6300	0.3196
0.3325	39.23	6400	0.3214
0.3321	39.85	6500	0.3190
0.3326	40.46	6600	0.3191
0.3323	41.07	6700	0.3215
0.3325	41.69	6800	0.3197
0.3325	42.3	6900	0.3198
0.3315	42.91	7000	0.3194
0.3317	43.52	7100	0.3196
0.3326	44.14	7200	0.3234
0.3304	44.75	7300	0.3196
0.3308	45.36	7400	0.3207
0.3313	45.98	7500	0.3182
0.3308	46.59	7600	0.3182
0.3305	47.2	7700	0.3188
0.3308	47.82	7800	0.3193
0.3313	48.43	7900	0.3199
0.3306	49.04	8000	0.3201
0.3307	49.66	8100	0.3187
0.3295	50.27	8200	0.3185
0.3298	50.88	8300	0.3190
0.3301	51.49	8400	0.3205
0.3299	52.11	8500	0.3202
0.3297	52.72	8600	0.3212
0.3302	53.33	8700	0.3206
0.3288	53.95	8800	0.3192
0.3286	54.56	8900	0.3189
0.3287	55.17	9000	0.3193
0.3302	55.79	9100	0.3191
0.328	56.4	9200	0.3196
0.3292	57.01	9300	0.3188
0.3288	57.62	9400	0.3175
0.3274	58.24	9500	0.3194
0.3289	58.85	9600	0.3191
0.3287	59.46	9700	0.3179
0.3293	60.08	9800	0.3208
0.3279	60.69	9900	0.3199
0.3282	61.3	10000	0.3193

Framework versions

• Transformers 4.38.2
• Pytorch 2.1.2
• Datasets 2.18.0
• Tokenizers 0.15.2