Hugging Face's logo

susmitabhatt
/
xlsr-big-kznnn

Automatic Speech Recognition
Transformers
TensorBoard
Safetensors
wav2vec2
Generated from Trainer
Inference Endpoints
Model card Files Metrics Community
Edit model card

xlsr-big-kznnn

This model is a fine-tuned version of facebook/wav2vec2-large-xlsr-53 on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 0.0000
  • Wer: 0.0559

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: 0.0004
  • train_batch_size: 8
  • eval_batch_size: 8
  • seed: 42
  • gradient_accumulation_steps: 2
  • total_train_batch_size: 16
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 132
  • num_epochs: 100
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Wer
2.1554 1.7167 200 0.9501 0.7058
0.6034 3.4335 400 0.0748 0.1296
0.1588 5.1502 600 0.0226 0.0679
0.0906 6.8670 800 0.0070 0.0569
0.0597 8.5837 1000 0.0181 0.0977
0.0523 10.3004 1200 0.0038 0.0547
0.0441 12.0172 1400 0.0028 0.0533
0.0362 13.7339 1600 0.0030 0.0593
0.0272 15.4506 1800 0.0055 0.0611
0.0286 17.1674 2000 0.0021 0.0565
0.0224 18.8841 2200 0.0036 0.0748
0.0227 20.6009 2400 0.0023 0.0543
0.0175 22.3176 2600 0.0075 0.0569
0.0172 24.0343 2800 0.0031 0.0547
0.0226 25.7511 3000 0.0027 0.0593
0.0132 27.4678 3200 0.0012 0.0535
0.0179 29.1845 3400 0.0022 0.0565
0.0149 30.9013 3600 0.0014 0.0531
0.0141 32.6180 3800 0.0010 0.0533
0.0146 34.3348 4000 0.0020 0.0565
0.0143 36.0515 4200 0.0002 0.0605
0.0117 37.7682 4400 0.0047 0.0567
0.0138 39.4850 4600 0.0011 0.0561
0.0095 41.2017 4800 0.0002 0.0738
0.0096 42.9185 5000 0.0001 0.0697
0.0083 44.6352 5200 0.0019 0.0601
0.0101 46.3519 5400 0.0010 0.0695
0.0078 48.0687 5600 0.0002 0.0571
0.0105 49.7854 5800 0.0002 0.0537
0.0065 51.5021 6000 0.0038 0.0681
0.0083 53.2189 6200 0.0000 0.0645
0.0065 54.9356 6400 0.0002 0.0543
0.0076 56.6524 6600 0.0005 0.0609
0.0055 58.3691 6800 0.0015 0.0557
0.0054 60.0858 7000 0.0001 0.0529
0.0078 61.8026 7200 0.0002 0.0525
0.0059 63.5193 7400 0.0002 0.0531
0.0051 65.2361 7600 0.0000 0.0535
0.0047 66.9528 7800 0.0000 0.0537
0.0045 68.6695 8000 0.0000 0.0549
0.0044 70.3863 8200 0.0014 0.0599
0.0051 72.1030 8400 0.0000 0.0551
0.003 73.8197 8600 0.0003 0.0547
0.0028 75.5365 8800 0.0000 0.0518
0.0027 77.2532 9000 0.0000 0.0520
0.0024 78.9700 9200 0.0000 0.0569
0.0022 80.6867 9400 0.0000 0.0565
0.0027 82.4034 9600 0.0000 0.0516
0.0018 84.1202 9800 0.0000 0.0520
0.0028 85.8369 10000 0.0000 0.0569
0.002 87.5536 10200 0.0000 0.0545
0.0014 89.2704 10400 0.0000 0.0535
0.0012 90.9871 10600 0.0000 0.0525
0.0011 92.7039 10800 0.0000 0.0531
0.0011 94.4206 11000 0.0000 0.0539
0.0009 96.1373 11200 0.0000 0.0555
0.0016 97.8541 11400 0.0000 0.0563
0.0011 99.5708 11600 0.0000 0.0559

Framework versions

  • Transformers 4.45.0.dev0
  • Pytorch 2.1.2
  • Datasets 2.20.0
  • Tokenizers 0.19.1
Downloads last month
8
Safetensors
Model size
315M params
Tensor type
F32
·
Inference Examples
Automatic Speech Recognition
This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Model tree for susmitabhatt/xlsr-big-kznnn

Base model

facebook/wav2vec2-large-xlsr-53
Finetuned
(210)
this model

Evaluation results

Metadata error: specify a dataset to view leaderboard