metadata
language:
- nl
license: apache-2.0
tags:
- automatic-speech-recognition
- mozilla-foundation/common_voice_8_0
- generated_from_trainer
- mt
- robust-speech-event
- model_for_talk
datasets:
- mozilla-foundation/common_voice_8_0
model-index:
- name: wav2vec2-large-xls-r-1b-cv8-mt-lm
results:
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: Common Voice 8
type: mozilla-foundation/common_voice_8_0
args: mt
metrics:
- name: Test WER
type: wer
value: 15.88
- name: Test CER
type: cer
value: 3.65
- task:
name: Automatic Speech Recognition
type: automatic-speech-recognition
dataset:
name: Robust Speech Event - Dev Data
type: speech-recognition-community-v2/dev_data
args: mt
metrics:
- name: Test WER
type: wer
value: null
- name: Test CER
type: cer
value: null
wav2vec2-large-xls-r-1b-cv8-mt-lm
This model is a fine-tuned version of wav2vec2-large-xls-r-1b-cv8-mt-lm on the common_voice 8 dataset. It achieves the following results on the test set:
- Loss: 0.2210
- Wer: 0.1974
Note that the above test results come from the original model without LM (language model) which can be found at https://huggingface.co/RuudVelo/wav2vec2-large-xls-r-1b-cv8-mt. The results with the LM model can be found on the right side of this model card.
Model description
Model RuudVelo/wav2vec2-large-xls-r-1b-cv8-mt which has been improved with a KenLM 3-gram.
Intended uses & limitations
More information needed
Training and evaluation data
Common Voice 8 mt dataset has been used for the model
Training procedure
Training hyperparameters
Parameters can be found in the run.sh file at https://huggingface.co/RuudVelo/wav2vec2-large-xls-r-1b-cv8-mt-lm
Framework versions
- Transformers 4.16.0.dev0
- Pytorch 1.10.1+cu102
- Datasets 1.18.3
- Tokenizers 0.11.0