metadata

library_name: transformers
license: apache-2.0
base_model: openai/whisper-medium
tags:
  - generated_from_trainer
datasets:
  - swagen
metrics:
  - wer
model-index:
  - name: whisper-medium-swagen-combined-30hrs-model
    results:
      - task:
          name: Automatic Speech Recognition
          type: automatic-speech-recognition
        dataset:
          name: swagen
          type: swagen
        metrics:
          - name: Wer
            type: wer
            value: 0.2234172077922078

whisper-medium-swagen-combined-30hrs-model

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

Loss: 0.3610
Wer: 0.2234

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: 42
gradient_accumulation_steps: 4
total_train_batch_size: 8
optimizer: Use 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
num_epochs: 30.0
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
2.7508	0.0828	200	0.8134	0.4877
1.8748	0.1656	400	0.6291	0.3898
1.6214	0.2484	600	0.5560	0.3431
1.559	0.3312	800	0.4968	0.2953
1.3616	0.4140	1000	0.4720	0.2872
1.3078	0.4967	1200	0.4577	0.2978
1.2579	0.5795	1400	0.4218	0.2758
1.214	0.6623	1600	0.4156	0.2654
1.0719	0.7451	1800	0.4005	0.2315
1.0432	0.8279	2000	0.3864	0.2433
0.9825	0.9107	2200	0.3743	0.2207
1.0952	0.9935	2400	0.3610	0.2234
0.6001	1.0766	2600	0.3888	0.2423
0.5491	1.1594	2800	0.3730	0.2265
0.6732	1.2422	3000	0.3702	0.2201

Framework versions

Transformers 4.47.1
Pytorch 2.5.1+cu124
Datasets 3.2.0
Tokenizers 0.21.0