Ukrainian STT model (with Language Model)

🇺🇦 Join Ukrainian Speech Recognition Community - https://t.me/speech_recognition_uk

⭐ See other Ukrainian models - https://github.com/egorsmkv/speech-recognition-uk

Have a look on an updated 300m model: https://huggingface.co/Yehor/wav2vec2-xls-r-300m-uk-with-small-lm
Have a look on a better model with more parameters: https://huggingface.co/Yehor/wav2vec2-xls-r-1b-uk-with-lm

This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m on the MOZILLA-FOUNDATION/COMMON_VOICE_7_0 - UK dataset. It achieves the following results on the evaluation set:

Loss: 0.3015
Wer: 0.3377
Cer: 0.0708

The above results present evaluation without the language model.

Model description

On 100 test example the model shows the following results:

Without LM:

WER: 0.2647
CER: 0.0469

With LM:

WER: 0.1568
CER: 0.0289

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 0.0003
train_batch_size: 8
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 20
total_train_batch_size: 160
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 500
num_epochs: 100.0
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
3.0255	7.93	500	2.5514	0.9921	0.9047
1.3809	15.86	1000	0.4065	0.5361	0.1201
1.2355	23.8	1500	0.3474	0.4618	0.1033
1.1956	31.74	2000	0.3617	0.4580	0.1005
1.1416	39.67	2500	0.3182	0.4074	0.0891
1.0996	47.61	3000	0.3166	0.3985	0.0875
1.0427	55.55	3500	0.3116	0.3835	0.0828
0.9961	63.49	4000	0.3137	0.3757	0.0807
0.9575	71.42	4500	0.2992	0.3632	0.0771
0.9154	79.36	5000	0.3015	0.3502	0.0740
0.8994	87.3	5500	0.3004	0.3425	0.0723
0.871	95.24	6000	0.3016	0.3394	0.0713

Framework versions

Transformers 4.16.0.dev0
Pytorch 1.10.1+cu102
Datasets 1.18.1.dev0
Tokenizers 0.11.0

Yehor
/

wav2vec2-xls-r-300m-uk-with-lm