./openai/whisper-medium.en-cit-do015-wd0-lr3e-06-FULL

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

• Loss: 0.6528
• Wer Ortho: 32.2429
• Wer: 22.4370

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-06
• train_batch_size: 4
• eval_batch_size: 8
• seed: 42
• distributed_type: multi-GPU
• gradient_accumulation_steps: 4
• total_train_batch_size: 16
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• training_steps: 500
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer Ortho	Wer
1.8096	0.4773	50	1.2178	42.1611	31.5966
1.1953	0.9547	100	0.9199	37.1498	27.2773
0.9212	1.4320	150	0.8408	34.7486	25.2605
0.8448	1.9093	200	0.7837	33.6001	24.5210
0.7174	2.3866	250	0.7344	32.5039	22.9076
0.6519	2.8640	300	0.7002	33.3391	23.4958
0.5866	3.3413	350	0.6802	32.2429	22.7395
0.5625	3.8186	400	0.6631	32.6083	22.8067
0.5207	4.2959	450	0.6548	32.6779	22.8908
0.5059	4.7733	500	0.6528	32.2429	22.4370

Framework versions

• Transformers 4.42.4
• Pytorch 1.13.1+cu117
• Datasets 2.20.0
• Tokenizers 0.19.1