whisper-large-v3-atcosim

This model is a fine-tuned version of openai/whisper-large-v3 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.0573
• Wer: 15.7807

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: 16
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• num_devices: 4
• total_train_batch_size: 64
• total_eval_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 250
• training_steps: 12500
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.0031	8.33	1000	0.0372	54.8342
0.0005	16.67	2000	0.0415	20.1519
0.0024	25.0	3000	0.0392	10.2102
0.0	33.33	4000	0.0469	18.6609
0.0	41.67	5000	0.0493	17.3180
0.0	50.0	6000	0.0511	16.8179
0.0	58.33	7000	0.0526	16.4753
0.0	66.67	8000	0.0538	16.5725
0.0	75.0	9000	0.0550	15.9983
0.0	83.33	10000	0.0560	15.7205
0.0	91.67	11000	0.0568	15.7159
0.0	100.0	12000	0.0573	15.7807

Framework versions

• Transformers 4.35.0
• Pytorch 2.0.1+cu117
• Datasets 2.12.0
• Tokenizers 0.14.1