wav2vec2-large-960h-lv60-self-dysarthria

This model is a fine-tuned version of facebook/wav2vec2-large-960h-lv60-self on the None dataset. It achieves the following results on the evaluation set:

  • Loss: 2.2385
  • Wer: 1.0

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: 0.0003
  • 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: 100
  • num_epochs: 30

Training results

Training Loss Epoch Step Validation Loss Wer
20.9703 3.2 200 3.5607 1.0
3.4216 6.4 400 3.3741 1.0
3.3062 9.6 600 3.1419 1.0
2.994 12.8 800 2.6835 1.0
2.7361 16.0 1000 2.4129 1.0
2.6005 19.2 1200 2.2973 1.0
2.5724 22.4 1400 2.2520 1.0
2.619 25.6 1600 2.2435 1.0
2.5278 28.8 1800 2.2385 1.0

Framework versions

  • Transformers 4.43.2
  • Pytorch 2.2.1+cu121
  • Datasets 2.20.0
  • Tokenizers 0.19.1
Downloads last month
28
Safetensors
Model size
315M params
Tensor type
BF16
·
Inference Examples
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 jmaczan/wav2vec2-large-960h-lv60-self-dysarthria

Finetuned
(5)
this model