whisper-large-v3-de-emodb-emotion-classification

This model is a fine-tuned version of openai/whisper-large-v3 on the Emo-DB dataset. It achieves the following results on the evaluation set:

  • Loss: 0.3724
  • Accuracy: 0.9439

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: 5e-05
  • train_batch_size: 2
  • eval_batch_size: 2
  • seed: 42
  • 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
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Accuracy
1.3351 1.0 214 1.1022 0.4953
0.2644 2.0 428 0.7572 0.7477
0.3796 3.0 642 1.0055 0.8131
0.0038 4.0 856 1.0754 0.8131
0.001 5.0 1070 0.5485 0.9159
0.001 6.0 1284 0.5881 0.8785
0.0007 7.0 1498 0.3376 0.9439
0.0006 8.0 1712 0.3592 0.9439
0.0006 9.0 1926 0.3695 0.9439
0.0004 10.0 2140 0.3724 0.9439

Framework versions

  • Transformers 4.45.0.dev0
  • Pytorch 2.4.0+cu121
  • Datasets 3.0.0
  • Tokenizers 0.19.1
Downloads last month
18
Safetensors
Model size
637M params
Tensor type
F32
ยท
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for Flocksserver/whisper-large-v3-de-emodb-emotion-classification

Finetuned
(599)
this model

Evaluation results