w2v-bert-hindi_voip

cdactvm/w2v-bert-hindi_voip

w2v-bert-hindi_voip is a automatic speech recognition model from cdactvm. Use it when you need speech turned into text. It is set up for transformers. The card lists the license as mit.

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

Downloads · 30 days

26% of all-time downloads

All-time downloads

Public

Parameters

606M

72.7 GB on disk

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.safetensors2.4 GB · 100%

At a glance

Task: Automatic Speech Recognition
Library: transformers
License: mit
Model type: wav2vec2-bert
Access: Public
Created: May 16, 2025
Updated: May 18, 2025
SHA: 9d4b5bb4

Base models

facebook/w2v-bert-2.0

Model card

From the Hugging Face model README

w2v-bert-hindi_voip

This model is a fine-tuned version of facebook/w2v-bert-2.0 on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.1299
Wer: 0.0737

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: 5e-05
train_batch_size: 2
eval_batch_size: 2
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 8
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: 500
training_steps: 60000
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.9895	0.1439	2000	0.8051	0.4624
0.8084	0.2878	4000	0.6416	0.3877
0.6948	0.4317	6000	0.5950	0.3660
0.6599	0.5756	8000	0.5228	0.3330
0.564	0.7195	10000	0.4980	0.3001
0.5112	0.8634	12000	0.4413	0.2928
0.4827	1.0073	14000	0.3999	0.2646
0.4198	1.1512	16000	0.3734	0.2438
0.3831	1.2951	18000	0.3779	0.2365
0.3699	1.4390	20000	0.3223	0.2144
0.34	1.5828	22000	0.3074	0.1973
0.3157	1.7267	24000	0.2809	0.1922
0.3041	1.8706	26000	0.2538	0.1751
0.2642	2.0145	28000	0.2620	0.1687
0.2296	2.1584	30000	0.2355	0.1521
0.2099	2.3023	32000	0.2279	0.1454
0.2037	2.4462	34000	0.2235	0.1405
0.1905	2.5901	36000	0.1988	0.1244
0.1876	2.7340	38000	0.1919	0.1227
0.1739	2.8779	40000	0.1865	0.1153
0.1387	3.0218	42000	0.1831	0.1077
0.1261	3.1657	44000	0.1760	0.1007
0.1146	3.3096	46000	0.1597	0.0945
0.1142	3.4535	48000	0.1602	0.0936
0.0998	3.5974	50000	0.1472	0.0861
0.102	3.7413	52000	0.1413	0.0823
0.0985	3.8852	54000	0.1376	0.0783
0.08	4.0291	56000	0.1338	0.0765
0.072	4.1730	58000	0.1319	0.0743
0.0677	4.3169	60000	0.1299	0.0737

Framework versions

Transformers 4.51.3
Pytorch 2.7.0+cu126
Datasets 3.6.0
Tokenizers 0.21.1

Task: Automatic Speech Recognition
Library: transformers
Type: wav2vec2-bert
License: mit
Created: May 16, 2025
Updated: May 18, 2025