metadata

language:
  - nl
license: apache-2.0
base_model: openai/whisper-large-v2
tags:
  - generated_from_trainer
metrics:
  - wer
model-index:
  - name: Whisper Large V2
    results: []

Whisper Large V2

This model is a fine-tuned version of openai/whisper-large-v2 on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.4586
Wer: 22.0846

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: 12
eval_batch_size: 8
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 20
num_epochs: 5

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.7137	0.1852	30	0.4795	35.1951
0.4678	0.3704	60	0.4433	43.9372
0.4255	0.5556	90	0.4112	36.2069
0.4164	0.7407	120	0.3886	29.2672
0.4173	0.9259	150	0.3784	30.1063
0.2763	1.1111	180	0.3815	29.2618
0.2067	1.2963	210	0.3685	25.0445
0.2142	1.4815	240	0.3779	32.5320
0.2079	1.6667	270	0.3749	22.7619
0.1928	1.8519	300	0.3703	24.0840
0.1812	2.0370	330	0.3788	21.3183
0.0992	2.2222	360	0.3961	20.3659
0.1028	2.4074	390	0.3885	26.0914
0.0992	2.5926	420	0.3888	22.8401
0.0957	2.7778	450	0.3771	26.2965
0.1015	2.9630	480	0.3729	24.6641
0.0607	3.1481	510	0.4192	21.8472
0.05	3.3333	540	0.4098	20.6276
0.0463	3.5185	570	0.4043	23.7035
0.0404	3.7037	600	0.4159	21.9524
0.0359	3.8889	630	0.4163	23.7116
0.0303	4.0741	660	0.4252	22.9588
0.014	4.2593	690	0.4608	22.4030
0.0152	4.4444	720	0.4586	22.7106
0.0142	4.6296	750	0.4547	22.7430
0.0143	4.8148	780	0.4591	21.9875
0.0138	5.0	810	0.4586	22.0846

Framework versions

Transformers 4.44.0.dev0
Pytorch 2.1.0+cu121
Datasets 2.20.0
Tokenizers 0.19.1