wav2vec2-tiny-fp16-demo

yenpolin/wav2vec2-tiny-fp16-demo

wav2vec2-tiny-fp16-demo is a automatic speech recognition model from yenpolin. Use it when you need speech turned into text. It is set up for transformers.

should probably proofread and complete it, then remove this comment. --

Downloads · 30 days

12% of all-time downloads

All-time downloads

177

Public

Repo size

15.2 GB

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.bin30.8 MB · 100%

At a glance

Task: Automatic Speech Recognition
Library: transformers
Model type: wav2vec2
Access: Public
Created: Mar 7, 2023
Updated: Mar 13, 2023
SHA: 6123b13c

Model card

From the Hugging Face model README

wav2vec2-tiny-fp16-demo

This model is a fine-tuned version of yenpolin/wav2vec2-tiny on the DNA_R9.4.1 - NA dataset. It achieves the following results on the evaluation set:

Loss: 0.5995
Mean Acc: 51.7421
Median Acc: 56.8345

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: 0.0003
train_batch_size: 400
eval_batch_size: 800
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 500
num_epochs: 100.0
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Mean Acc	Median Acc
1.1754	1.0	1250	1.0956	1.8729	0.0
1.0286	2.0	2500	0.9805	12.8456	0.0
0.921	3.0	3750	0.9232	12.3399	0.0
0.8819	4.0	5000	0.8857	18.0811	0.0
0.8574	5.0	6250	0.8691	17.4596	0.0
0.8419	6.0	7500	0.8578	25.5205	0.0
0.8291	7.0	8750	0.8441	23.7467	0.0
0.8167	8.0	10000	0.8346	22.1622	0.0
0.801	9.0	11250	0.8108	31.3152	53.4606
0.7768	10.0	12500	0.7868	30.8526	53.2143
0.7533	11.0	13750	0.7642	34.3817	53.9589
0.7385	12.0	15000	0.7637	39.9523	54.6539
0.7273	13.0	16250	0.7412	38.2095	54.5455
0.7103	14.0	17500	0.7212	39.3217	54.6032
0.6876	15.0	18750	0.7128	33.8479	53.8226
0.6744	16.0	20000	0.6916	36.8302	54.0541
0.6664	17.0	21250	0.6825	39.4562	54.4928
0.6598	18.0	22500	0.6753	36.5260	53.9130
0.654	19.0	23750	0.6748	33.4539	53.6058
0.6492	20.0	25000	0.6702	37.7485	54.1209
0.6457	21.0	26250	0.6694	40.0633	54.4343
0.6413	22.0	27500	0.6609	36.9515	53.9062
0.6377	23.0	28750	0.6576	39.6921	54.2857
0.6341	24.0	30000	0.6527	41.3149	54.5455
0.6312	25.0	31250	0.6516	39.5467	54.2587
0.6279	26.0	32500	0.6487	38.4877	54.0541
0.6246	27.0	33750	0.6444	43.1149	54.7619
0.6219	28.0	35000	0.6439	40.7786	54.5151
0.6191	29.0	36250	0.6418	41.0059	54.5161
0.6167	30.0	37500	0.6359	42.2812	54.7425
0.6136	31.0	38750	0.6346	41.2515	54.5977
0.6111	32.0	40000	0.6332	41.7222	54.6667
0.6087	33.0	41250	0.6333	41.3847	54.5455
0.6062	34.0	42500	0.6279	43.7450	55.0000
0.6042	35.0	43750	0.6269	42.5222	54.7619
0.6021	36.0	45000	0.6288	43.8255	54.9133
0.6004	37.0	46250	0.6292	43.5378	55.0186
0.5987	38.0	47500	0.6233	43.8364	55.0000
0.5969	39.0	48750	0.6213	44.3750	55.1155
0.595	40.0	50000	0.6207	40.1240	54.5732
0.5934	41.0	51250	0.6198	44.3429	55.0885
0.5914	42.0	52500	0.6178	44.0396	55.1181
0.5895	43.0	53750	0.6155	45.7166	55.4017
0.5874	44.0	55000	0.6159	46.3816	55.4745
0.5853	45.0	56250	0.6136	45.7973	55.4810
0.5834	46.0	57500	0.6142	44.1374	55.2
0.5819	47.0	58750	0.6204	46.3006	55.4945
0.5802	48.0	60000	0.6093	47.6637	55.7276
0.5785	49.0	61250	0.6098	46.7728	55.625
0.5771	50.0	62500	0.6075	47.4424	55.7423
0.5751	51.0	63750	0.6066	46.8072	55.7377
0.5734	52.0	65000	0.6053	48.7600	56.0109
0.5719	53.0	66250	0.6048	48.8645	55.9557
0.5705	54.0	67500	0.6030	48.4536	55.9557
0.5693	55.0	68750	0.6046	49.0295	56.0870
0.5677	56.0	70000	0.6034	47.9864	55.9767
0.5665	57.0	71250	0.6033	48.4370	56.0241
0.5653	58.0	72500	0.6016	49.2084	56.1069
0.5635	59.0	73750	0.6021	49.6844	56.2648
0.5621	60.0	75000	0.6001	50.2077	56.3218
0.5607	61.0	76250	0.6007	49.1341	56.1753
0.5592	62.0	77500	0.6011	48.9626	56.2016
0.5582	63.0	78750	0.6008	49.6105	56.2982
0.557	64.0	80000	0.6012	49.5089	56.2814
0.5557	65.0	81250	0.5994	49.9531	56.3246
0.5542	66.0	82500	0.6003	50.2286	56.4706
0.5533	67.0	83750	0.5990	49.9327	56.3725
0.5521	68.0	85000	0.5995	50.1928	56.4516
0.5511	69.0	86250	0.5979	50.2263	56.4444
0.55	70.0	87500	0.6005	50.4875	56.5341
0.5484	71.0	88750	0.5984	50.0615	56.4626
0.5475	72.0	90000	0.5991	50.4546	56.5598
0.5467	73.0	91250	0.5972	50.8276	56.5728
0.5458	74.0	92500	0.5987	51.1359	56.5934
0.5445	75.0	93750	0.5986	51.0738	56.6474
0.5439	76.0	95000	0.5991	50.7110	56.6102
0.5431	77.0	96250	0.5986	51.0240	56.6667
0.5423	78.0	97500	0.5992	51.0228	56.6901
0.5413	79.0	98750	0.5992	50.8097	56.6343
0.5407	80.0	100000	0.5993	51.2619	56.7251
0.5399	81.0	101250	0.5975	51.4024	56.7282
0.5393	82.0	102500	0.5982	51.2136	56.7358
0.5386	83.0	103750	0.5983	51.4384	56.7376
0.538	84.0	105000	0.5988	51.6233	56.7568
0.5374	85.0	106250	0.5983	51.4155	56.7867
0.5369	86.0	107500	0.5987	51.4146	56.7568
0.5365	87.0	108750	0.5995	51.4874	56.8
0.5361	88.0	110000	0.5994	51.5007	56.7775
0.5357	89.0	111250	0.5991	51.5986	56.7901
0.5353	90.0	112500	0.5990	51.4792	56.7742
0.535	91.0	113750	0.5993	51.8824	56.8528
0.5348	92.0	115000	0.5993	51.6289	56.8116
0.5346	93.0	116250	0.5993	51.6716	56.8254
0.5344	94.0	117500	0.5991	51.5984	56.8
0.5342	95.0	118750	0.5995	51.8793	56.8245
0.5341	96.0	120000	0.5994	51.7306	56.8254
0.534	97.0	121250	0.5994	51.7991	56.8306
0.5339	98.0	122500	0.5993	51.8468	56.8233
0.5339	99.0	123750	0.5994	51.8060	56.8389
0.5338	100.0	125000	0.5995	51.7421	56.8345

Framework versions

Transformers 4.26.0
Pytorch 1.12.1
Datasets 2.9.0
Tokenizers 0.13.2

Task: Automatic Speech Recognition
Library: transformers
Type: wav2vec2
Created: Mar 7, 2023
Updated: Mar 13, 2023