README.md

metadata

base_model: Shehryar718/URDU-ASR
tags:
  - generated_from_trainer
datasets:
  - common_voice_13_0
metrics:
  - wer
model-index:
  - name: URDU-ASR-25-EPOCH
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: common_voice_13_0
          type: common_voice_13_0
          config: ur
          split: test
          args: ur
        metrics:
          - name: Wer
            type: wer
            value: 0.47599520290920344

URDU-ASR-25-EPOCH

This model is a fine-tuned version of Shehryar718/URDU-ASR on the common_voice_13_0 dataset. It achieves the following results on the evaluation set:

• Loss: 0.6782
• Wer: 0.4760
• Cer: 0.1986

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: 7.5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.99) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 25
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Cer
5.3707	1.0	341	1.3583	0.8266	0.3484
0.4814	2.0	683	0.7213	0.5187	0.2196
0.2821	3.0	1024	0.6354	0.4917	0.2066
0.2368	4.0	1366	0.6730	0.5122	0.2181
0.2105	5.0	1707	0.6430	0.4871	0.2076
0.1965	6.0	2049	0.6397	0.4902	0.2136
0.1879	7.0	2390	0.6397	0.4698	0.1951
0.1743	8.0	2732	0.6636	0.4739	0.1996
0.1632	9.0	3073	0.6752	0.4782	0.1996
0.1572	10.0	3415	0.6859	0.4874	0.2072
0.1586	11.0	3756	0.6761	0.4844	0.2069
0.1595	12.0	4098	0.6846	0.4746	0.1959
0.1534	13.0	4439	0.6750	0.4830	0.2034
0.16	14.0	4781	0.6653	0.4826	0.2038
0.1752	15.0	5122	0.6536	0.4727	0.1946
0.1739	16.0	5464	0.6753	0.4738	0.1912
0.1709	17.0	5805	0.6600	0.4730	0.1996
0.1676	18.0	6147	0.6691	0.4678	0.1919
0.1636	19.0	6488	0.6638	0.4772	0.1990
0.1593	20.0	6830	0.6787	0.4764	0.1976
0.1588	21.0	7171	0.6699	0.4772	0.1974
0.1525	22.0	7513	0.6827	0.4738	0.1962
0.1554	23.0	7854	0.6740	0.4736	0.1970
0.1522	24.0	8196	0.6791	0.4768	0.1989
0.1502	24.96	8525	0.6782	0.4760	0.1986

Framework versions

• Transformers 4.35.0
• Pytorch 2.1.0+cu121
• Datasets 2.14.4
• Tokenizers 0.14.1