wav2vec2-large-xlsr-300m-nepali

This model is a fine-tuned version of facebook/wav2vec2-xls-r-300m

The dataset used to train are :

1. OpenSLR-54 Corpus
2. External Data

For evaluation on publicly available datasets, can use OpenSLR-43 corpus - Model is not trained on this data which is also available on HuggingFace as train-set where it achieves 27% WER and 8.3% CER with 5 gram language model.

Script to Evaluate the Model on OpenSLR-43 train set :

python3 eval.py --model_id prajin/wav2vec2-large-xlsr-300m-nepali --dataset openslr --config SLR43  --split train --log_outputs

Below evaluation result is the evaluation on separated 10000 samples from total training dataset.

It achieves the following results on the evaluation set Without using Language Model :

• Loss: 0.2625
• Wer: 0.3426

With Language model ( 5 gram )

• Wer: 0.2502

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 6e-05
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 1000
• num_epochs: 1
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.3468	0.04	400	0.2624	0.3479
0.2792	0.08	800	0.2696	0.3490
0.245	0.12	1200	0.2751	0.3502
0.2453	0.16	1600	0.2754	0.3523
0.2332	0.2	2000	0.2779	0.3517
0.2321	0.24	2400	0.2775	0.3528
0.2708	0.28	2800	0.2764	0.3533
0.2709	0.32	3200	0.2723	0.3544
0.2715	0.36	3600	0.2739	0.3545
0.2732	0.4	4000	0.2707	0.3498
0.2643	0.44	4400	0.2696	0.3499
0.2682	0.47	4800	0.2672	0.3492
0.2687	0.51	5200	0.2644	0.3474
0.269	0.55	5600	0.2619	0.3502
0.2675	0.59	6000	0.2606	0.3477
0.2656	0.63	6400	0.2597	0.3463
0.2667	0.67	6800	0.2607	0.3458
0.2639	0.71	7200	0.2601	0.3480
0.2631	0.75	7600	0.2582	0.3447
0.2589	0.79	8000	0.2577	0.3438
0.2554	0.83	8400	0.2557	0.3439
0.2687	0.87	8800	0.2546	0.3438
0.2574	0.91	9200	0.2537	0.3434
0.2623	0.95	9600	0.2530	0.3433
0.2675	0.99	10000	0.2530	0.3426

Framework versions

• Transformers 4.16.0.dev0
• Pytorch 1.10.1+cu102
• Datasets 1.17.1.dev0
• Tokenizers 0.11.0