kbd-asr-colab2

MatricariaV/kbd-asr-colab2

kbd-asr-colab2 is a automatic speech recognition model from MatricariaV. 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

4% of all-time downloads

All-time downloads

689

Public

Parameters

965M

96.5 GB on disk

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.safetensors3.9 GB · 100%

At a glance

Task: Automatic Speech Recognition
Library: transformers
License: cc-by-nc-4.0
Model type: wav2vec2
Access: Public
Created: Feb 24, 2025
Updated: Feb 24, 2025
SHA: 1ce5e74c

Base models

facebook/mms-1b-all

Spaces

MatricariaV/asr-kabardian-demo

Model card

From the Hugging Face model README

kbd-asr-colab2

This model is a fine-tuned version of facebook/mms-1b-all on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.1829
Wer: 0.3182

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: 0.001
train_batch_size: 8
eval_batch_size: 8
seed: 42
optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 100
num_epochs: 5
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.7968	0.1871	200	0.4521	0.7382
0.6536	0.3742	400	0.3965	0.7499
0.5989	0.5613	600	0.3850	0.7644
0.5212	0.7484	800	0.3473	0.6373
0.5649	0.9355	1000	0.2951	0.5633
0.4996	1.1225	1200	0.2771	0.4806
0.5247	1.3096	1400	0.2553	0.4592
0.4559	1.4967	1600	0.2537	0.4396
0.5153	1.6838	1800	0.2479	0.4364
0.4622	1.8709	2000	0.2363	0.4251
0.4533	2.0580	2200	0.2280	0.4091
0.4529	2.2451	2400	0.2182	0.4066
0.4453	2.4322	2600	0.2191	0.3902
0.4339	2.6193	2800	0.2135	0.3815
0.4179	2.8064	3000	0.2151	0.3906
0.3983	2.9935	3200	0.2059	0.3725
0.3868	3.1805	3400	0.2058	0.3614
0.3804	3.3676	3600	0.1959	0.3558
0.3934	3.5547	3800	0.2015	0.3555
0.3693	3.7418	4000	0.1995	0.3587
0.3678	3.9289	4200	0.2017	0.3876
0.4097	4.1160	4400	0.1893	0.3355
0.34	4.3031	4600	0.1877	0.3291
0.3553	4.4902	4800	0.1857	0.3222
0.3133	4.6773	5000	0.1839	0.3224
0.3077	4.8644	5200	0.1829	0.3182

Framework versions

Transformers 4.50.0.dev0
Pytorch 2.5.1+cu121
Datasets 3.3.2
Tokenizers 0.21.0

Task: Automatic Speech Recognition
Library: transformers
Type: wav2vec2
License: cc-by-nc-4.0
Created: Feb 24, 2025
Updated: Feb 24, 2025