Edit model card

hatespeech_wav2vec2

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

  • Loss: 0.6562
  • Accuracy: 0.6216
  • Recall: 0.7853
  • Precision: 0.5990
  • F1: 0.6796

It achieves the following results on the test set:

  • Loss: 0.6597
  • Accuracy: 0.6192
  • Recall: 0.7822
  • Precision: 0.5944
  • F1: 0.6755

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: 4e-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 Recall Precision F1
No log 0.9935 77 0.6871 0.5430 0.9021 0.5311 0.6686
0.6899 2.0 155 0.6779 0.5647 0.9021 0.5448 0.6793
0.6761 2.9935 232 0.6649 0.5934 0.5541 0.6131 0.5821
0.6607 4.0 310 0.6550 0.6289 0.6504 0.6334 0.6417
0.6607 4.9935 387 0.6562 0.6216 0.7853 0.5990 0.6796
0.6403 6.0 465 0.6578 0.6357 0.6969 0.6298 0.6617
0.6129 6.9935 542 0.6623 0.6313 0.7277 0.6184 0.6686
0.6024 8.0 620 0.6745 0.6345 0.7490 0.6174 0.6769
0.5779 8.9935 697 0.6807 0.6406 0.6567 0.6460 0.6513
0.5779 9.9355 770 0.6798 0.6337 0.6993 0.6270 0.6612

Framework versions

  • Transformers 4.40.2
  • Pytorch 2.2.1+cu121
  • Datasets 2.19.1
  • Tokenizers 0.19.1
Downloads last month
2
Safetensors
Model size
94.6M params
Tensor type
F32
·

Finetuned from

Dataset used to train DL-Project/hatespeech_wav2vec2