Edit model card

wav2vec_final_output

This model is a fine-tuned version of facebook/wav2vec2-base on the speech_commands dataset. It achieves the following results on the evaluation set:

  • Loss: 0.4410
  • Accuracy: 0.9018

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: 3e-05
  • train_batch_size: 32
  • eval_batch_size: 32
  • seed: 42
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 128
  • optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_ratio: 0.1
  • num_epochs: 10

Training results

Training Loss Epoch Step Validation Loss Accuracy
1.4588 1.0 663 1.2309 0.8763
0.6109 2.0 1326 0.5745 0.8920
0.4153 3.0 1989 0.4884 0.8953
0.3227 4.0 2652 0.4574 0.8980
0.2806 5.0 3315 0.4412 0.8994
0.207 6.0 3978 0.4403 0.9014
0.2226 7.0 4641 0.4479 0.8998
0.2577 8.0 5304 0.4421 0.9014
0.2188 9.0 5967 0.4408 0.9016
0.2082 10.0 6630 0.4410 0.9018

Framework versions

  • Transformers 4.34.1
  • Pytorch 2.1.0+cu118
  • Datasets 2.14.6
  • Tokenizers 0.14.1
Downloads last month
1
Safetensors
Model size
86.2M params
Tensor type
F32
·

Finetuned from

Dataset used to train moonseok/wav2vec_final_output

Evaluation results