wav2vec2-large-mms-1b-aft-bxk

robertp408/wav2vec2-large-mms-1b-aft-bxk

wav2vec2-large-mms-1b-aft-bxk is a automatic speech recognition model from robertp408. 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

236

Public

Parameters

965M

162 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: Sep 28, 2025
Updated: Oct 7, 2025
SHA: c3146c62

Base models

facebook/mms-1b-all

Datasets

audiofolder

Model card

From the Hugging Face model README

wav2vec2-large-mms-1b-aft-bxk

This model is a fine-tuned version of facebook/mms-1b-all on the audiofolder dataset. It achieves the following results on the evaluation set:

Loss: 0.5621
Wer: 0.5131

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: 2
eval_batch_size: 8
seed: 42
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED 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
7.9796	0.1218	100	4.5240	1.0
3.7025	0.2436	200	3.5923	0.9996
3.1866	0.3654	300	3.5363	1.0009
3.6159	0.4872	400	3.3896	1.0110
2.5623	0.6090	500	1.0336	0.8341
0.7145	0.7308	600	0.7236	0.6466
0.6345	0.8526	700	0.6972	0.6094
0.5631	0.9744	800	0.7104	0.6126
0.5549	1.0962	900	0.6706	0.5859
0.538	1.2180	1000	0.6259	0.5737
0.522	1.3398	1100	0.6353	0.5655
0.5596	1.4616	1200	0.6187	0.5602
0.5453	1.5834	1300	0.6135	0.5517
0.5044	1.7052	1400	0.6166	0.5428
0.4995	1.8270	1500	0.6083	0.5535
0.5385	1.9488	1600	0.5863	0.5430
0.5391	2.0706	1700	0.6098	0.5405
0.4674	2.1924	1800	0.6193	0.5487
0.4865	2.3143	1900	0.5871	0.5473
0.4584	2.4361	2000	0.6001	0.5563
0.444	2.5579	2100	0.5956	0.5420
0.4609	2.6797	2200	0.5955	0.5312
0.5002	2.8015	2300	0.5920	0.5273
0.5037	2.9233	2400	0.5783	0.5367
0.4699	3.0451	2500	0.5740	0.5269
0.4855	3.1669	2600	0.5859	0.5189
0.4253	3.2887	2700	0.5913	0.5205
0.4612	3.4105	2800	0.5729	0.5230
0.4461	3.5323	2900	0.5860	0.5198
0.4609	3.6541	3000	0.5650	0.5235
0.4501	3.7759	3100	0.5755	0.5239
0.4104	3.8977	3200	0.5723	0.5212
0.4309	4.0195	3300	0.5844	0.5145
0.4354	4.1413	3400	0.5714	0.5175
0.4077	4.2631	3500	0.5702	0.5163
0.4198	4.3849	3600	0.5658	0.5112
0.3993	4.5067	3700	0.5676	0.5174
0.4637	4.6285	3800	0.5660	0.5106
0.4198	4.7503	3900	0.5632	0.5108
0.4363	4.8721	4000	0.5625	0.5115
0.4041	4.9939	4100	0.5621	0.5131

Framework versions

Transformers 4.57.0
Pytorch 2.8.0+cu128
Datasets 4.1.1
Tokenizers 0.22.1

Task: Automatic Speech Recognition
Library: transformers
Type: wav2vec2
License: cc-by-nc-4.0
Created: Sep 28, 2025
Updated: Oct 7, 2025