whisper-medium-bigcgen-combined-5hrs-62

This model is a fine-tuned version of openai/whisper-medium on the bigcgen dataset. It achieves the following results on the evaluation set:

  • Loss: 0.7209
  • Wer: 0.5565

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: 1e-05
  • train_batch_size: 2
  • eval_batch_size: 2
  • seed: 62
  • gradient_accumulation_steps: 4
  • total_train_batch_size: 8
  • optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
  • lr_scheduler_type: linear
  • lr_scheduler_warmup_steps: 500
  • training_steps: 5000
  • mixed_precision_training: Native AMP

Training results

Training Loss Epoch Step Validation Loss Wer
1.0959 0.6079 200 0.9655 0.6610
0.6562 1.2158 400 0.8052 0.5699
0.6063 1.8237 600 0.7209 0.5565
0.402 2.4316 800 0.7347 0.5566
0.3066 3.0395 1000 0.7320 0.5467
0.2262 3.6474 1200 0.7329 0.5896
0.117 4.2553 1400 0.7819 0.5211

Framework versions

  • Transformers 4.53.0.dev0
  • Pytorch 2.6.0+cu124
  • Datasets 3.6.0
  • Tokenizers 0.21.0
Downloads last month
17
Safetensors
Model size
764M 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 csikasote/whisper-medium-bigcgen-combined-5hrs-62

Finetuned
(700)
this model

Evaluation results