AI4Bharat / IndicWav2Vec

Pretraining, fine-tuning and evaluation scripts for Indic-Wav2Vec2
https://indicnlp.ai4bharat.org/indicwav2vec
MIT License
82 stars 28 forks source link

What is the difference between different language model names? #8

Closed raotnameh closed 2 years ago

raotnameh commented 2 years ago

Hi Team, @gowtham1997 What is the difference? small_2 and small or comb?

bash infer_auto.sh 0 "hindi_mucs" "large" "mucs_small_2" 2 -4 1024 &\
bash infer_auto.sh 0 "hindi_mucs" "large" "mucs_large_2" 2 -4 1024 &\
bash infer_auto.sh 0 "hindi_mucs" "large" "comb" 2 -4 1024 &\
bash infer_auto.sh 0 "hindi_mucs" "large" "mucs_small" 2 -4 1024 &\
bash infer_auto.sh 0 "hindi_mucs" "large" "mucs_large" 2 -4 1024 &\
RamanHacks commented 2 years ago

Hi @raotnameh , the language models differ in terms of the data/lexicon used for training them. For more info, please refer to Table 3 of our paper: small/small_2 corresponds to the M4 row large/large_2 corresponds to the M5 row comb corresponds to the M6 row (augmented_lexicon)