models_sv_eric_1

niclas/models_sv_eric_1

models_sv_eric_1 is a automatic speech recognition model from niclas. Use it when you need speech turned into text. It is set up for transformers. The card lists the license as cc-by-nc-4.0.

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

Downloads · 30 days

2% of all-time downloads

All-time downloads

283

Public

Repo size

34.1 GB

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.bin1.3 GB · 100%

At a glance

Task: Automatic Speech Recognition
Library: transformers
License: cc-by-nc-4.0
Model type: wav2vec2
Access: Public
Created: Aug 23, 2022
Updated: Aug 23, 2022
SHA: ed33882d

Model card

From the Hugging Face model README

models_sv_eric_1

This model is a fine-tuned version of facebook/wav2vec2-large-100k-voxpopuli on the None dataset. It achieves the following results on the evaluation set:

Loss: 2.1340
Wer: 0.6241

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: 4
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 2
total_train_batch_size: 8
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 1000
num_epochs: 300

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
27.2483	5.81	250	12.8968	1.0
5.3813	11.63	500	3.7635	1.0
3.1776	17.44	750	3.1586	1.0
3.0849	23.26	1000	3.1336	1.0
3.0351	29.07	1250	3.0069	1.0
2.5591	34.88	1500	1.8101	0.9735
1.4236	40.7	1750	1.3666	0.8120
0.9233	46.51	2000	1.3338	0.7470
0.6594	52.33	2250	1.4020	0.7060
0.5056	58.14	2500	1.3793	0.7036
0.4135	63.95	2750	1.3789	0.6988
0.3521	69.77	3000	1.4288	0.6795
0.2728	75.58	3250	1.4819	0.6554
0.2419	81.4	3500	1.5370	0.6602
0.2288	87.21	3750	1.4477	0.6265
0.2009	93.02	4000	1.5387	0.6602
0.1773	98.84	4250	1.6789	0.6723
0.1701	104.65	4500	1.6322	0.6361
0.1562	110.47	4750	1.5988	0.6554
0.1433	116.28	5000	1.7502	0.6458
0.1373	122.09	5250	1.7735	0.6217
0.1186	127.91	5500	1.7193	0.6193
0.1127	133.72	5750	1.8742	0.6410
0.113	139.53	6000	1.8339	0.6337
0.1106	145.35	6250	1.7486	0.6289
0.0955	151.16	6500	1.7455	0.6361
0.0934	156.98	6750	1.8922	0.6361
0.0873	162.79	7000	2.0495	0.6530
0.0863	168.6	7250	1.8438	0.6361
0.0901	174.42	7500	2.0441	0.6289
0.0749	180.23	7750	2.0112	0.6265
0.0887	186.05	8000	2.0684	0.6554
0.074	191.86	8250	2.0821	0.6265
0.0714	197.67	8500	2.0790	0.6313
0.0638	203.49	8750	2.0158	0.6072
0.0633	209.3	9000	2.0423	0.6386
0.0621	215.12	9250	2.0013	0.6241
0.0616	220.93	9500	1.9567	0.6386
0.0627	226.74	9750	2.0302	0.6361
0.0604	232.56	10000	2.0424	0.6096
0.0551	238.37	10250	2.0238	0.6096
0.0559	244.19	10500	2.0207	0.6361
0.0587	250.0	10750	2.0818	0.6361
0.0508	255.81	11000	2.1106	0.6289
0.0494	261.63	11250	2.1194	0.6434
0.0576	267.44	11500	2.0752	0.6410
0.0521	273.26	11750	2.1455	0.6361
0.0479	279.07	12000	2.1583	0.6337
0.0501	284.88	12250	2.1400	0.6386
0.0447	290.7	12500	2.1440	0.6265
0.0455	296.51	12750	2.1340	0.6241

Framework versions

Transformers 4.11.3
Pytorch 1.9.0
Datasets 1.13.3
Tokenizers 0.10.3

Task: Automatic Speech Recognition
Library: transformers
Type: wav2vec2
License: cc-by-nc-4.0
Created: Aug 23, 2022
Updated: Aug 23, 2022