Wav2vec2-XLS-R-1B-RO Open-Source Speech Recognition Model - Free Support for Romanian Speech Recognition

Wav2vec2 Xls R 1b Ro

Developed by ubamba98

This model is an automatic speech recognition model fine-tuned on the Romanian Common Voice 7.0 dataset based on facebook/wav2vec2-xls-r-1b.

Speech Recognition

Transformers

OtherOpen Source License:Apache-2.0 #Romanian speech recognition #Large model fine-tuning #Low CER

Downloads 16

Release Time : 3/2/2022

Model Overview

A model for Romanian automatic speech recognition (ASR), fine-tuned based on the wav2vec2-xls-r-1b architecture, suitable for speech-to-text tasks.

Model Features

Large-scale pre-trained model fine-tuning

Fine-tuned on the 1-billion-parameter wav2vec2-xls-r model with powerful speech feature extraction capabilities

Romanian language optimization

Specifically optimized for Romanian speech recognition tasks

Multi-dataset evaluation

Comprehensively evaluated on Common Voice 7.0 and robust speech event datasets

Model Capabilities

Romanian speech recognition

Speech-to-text

Automatic speech recognition

Use Cases

Speech transcription

Romanian speech-to-text

Convert Romanian speech content into text

Achieved a WER of 47.7% on the Common Voice 7.0 test set

Voice assistants

Romanian voice command recognition

Basic recognition component for Romanian voice assistants

🚀 wav2vec2-xls-r-1b-ro

This model is a fine - tuned version of facebook/wav2vec2-xls-r-1b on the MOZILLA - FOUNDATION/COMMON_VOICE_7_0 - RO dataset. It is designed for automatic speech recognition and can effectively handle speech data in Romanian.

🚀 Quick Start

This model is a fine - tuned version of facebook/wav2vec2-xls-r-1b on the MOZILLA - FOUNDATION/COMMON_VOICE_7_0 - RO dataset. It achieves the following results on the evaluation set:

Loss: 0.1113
Wer: 0.4770
Cer: 0.0306

📚 Documentation

Model Information

Property	Details
Language	Romanian
License	Apache - 2.0
Tags	automatic - speech - recognition, generated_from_trainer, hf - asr - leaderboard, mozilla - foundation/common_voice_7_0, robust - speech - event
Datasets	mozilla - foundation/common_voice_7_0

Model Performance

The model has the following performance on different datasets:

Task	Dataset	Test WER
Automatic Speech Recognition	Common Voice 7.0	99.99
Automatic Speech Recognition	Robust Speech Event - Dev Data	99.98
Automatic Speech Recognition	Robust Speech Event - Test Data	99.99

Training Procedure

Training Hyperparameters

The following hyperparameters were used during training:

learning_rate: 3e - 05
train_batch_size: 8
eval_batch_size: 8
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon = 1e - 08
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 2000
num_epochs: 50.0
mixed_precision_training: Native AMP

Training Results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
0.7844	1.67	1500	0.3412	0.8600	0.0940
0.7272	3.34	3000	0.1926	0.6409	0.0527
0.6924	5.02	4500	0.1413	0.5722	0.0401
0.6327	6.69	6000	0.1252	0.5366	0.0371
0.6363	8.36	7500	0.1235	0.5741	0.0389
0.6238	10.03	9000	0.1180	0.5542	0.0362
0.6018	11.71	10500	0.1192	0.5694	0.0369
0.583	13.38	12000	0.1216	0.5772	0.0385
0.5643	15.05	13500	0.1195	0.5419	0.0371
0.5399	16.72	15000	0.1240	0.5224	0.0370
0.5529	18.39	16500	0.1174	0.5555	0.0367
0.5246	20.07	18000	0.1097	0.5047	0.0339
0.4936	21.74	19500	0.1225	0.5189	0.0382
0.4629	23.41	21000	0.1142	0.5047	0.0344
0.4463	25.08	22500	0.1168	0.4887	0.0339
0.4671	26.76	24000	0.1119	0.5073	0.0338
0.4359	28.43	25500	0.1206	0.5479	0.0363
0.4225	30.1	27000	0.1122	0.5170	0.0345
0.4038	31.77	28500	0.1159	0.5032	0.0343
0.4271	33.44	30000	0.1116	0.5126	0.0339
0.3867	35.12	31500	0.1101	0.4937	0.0327
0.3674	36.79	33000	0.1142	0.4940	0.0330
0.3607	38.46	34500	0.1106	0.5145	0.0327
0.3651	40.13	36000	0.1172	0.4921	0.0317
0.3268	41.81	37500	0.1093	0.4830	0.0310
0.3345	43.48	39000	0.1131	0.4760	0.0314
0.3236	45.15	40500	0.1132	0.4864	0.0317
0.312	46.82	42000	0.1124	0.4861	0.0315
0.3106	48.49	43500	0.1116	0.4745	0.0306

Framework Versions

Transformers 4.17.0.dev0
Pytorch 1.10.2+cu102
Datasets 1.18.2.dev0
Tokenizers 0.11.0

📄 License

This model is licensed under the Apache - 2.0 license.

Featured Recommended AI Models

Empowering the Future, Your AI Solution Knowledge Base

English 简体中文繁體中文にほんご