wav2vec2-xls-r-300m-cv7-turkish

Model description

This ASR model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on Turkish language.

Training and evaluation data

The following datasets were used for finetuning:

• Common Voice 7.0 TR All validated split except test split was used for training.
• MediaSpeech

Training procedure

To support both of the datasets above, custom pre-processing and loading steps was performed and wav2vec2-turkish repo was used for that purpose.

Training hyperparameters

The following hypermaters were used for finetuning:

• learning_rate 2e-4
• num_train_epochs 10
• warmup_steps 500
• freeze_feature_extractor
• mask_time_prob 0.1
• mask_feature_prob 0.05
• feat_proj_dropout 0.05
• attention_dropout 0.05
• final_dropout 0.05
• activation_dropout 0.05
• per_device_train_batch_size 8
• per_device_eval_batch_size 8
• gradient_accumulation_steps 8

Framework versions

• Transformers 4.16.0.dev0
• Pytorch 1.10.1
• Datasets 1.17.0
• Tokenizers 0.10.3

Language Model

N-gram language model is trained on a Turkish Wikipedia articles using KenLM and ngram-lm-wiki repo was used to generate arpa LM and convert it into binary format.

Evaluation Commands

Please install unicode_tr package before running evaluation. It is used for Turkish text processing.

1. To evaluate on mozilla-foundation/common_voice_7_0 with split test

python eval.py --model_id mpoyraz/wav2vec2-xls-r-300m-cv7-turkish --dataset mozilla-foundation/common_voice_7_0 --config tr --split test

2. To evaluate on speech-recognition-community-v2/dev_data

python eval.py --model_id mpoyraz/wav2vec2-xls-r-300m-cv7-turkish --dataset speech-recognition-community-v2/dev_data --config tr --split validation --chunk_length_s 5.0 --stride_length_s 1.0

Evaluation results:

Dataset	WER	CER
Common Voice 7 TR test split	8.62	2.26
Speech Recognition Community dev data	30.87	10.69