deepinfinityai/30_sentencesV1

Whisper_Large_30_sent_Model

This model is a fine-tuned version of openai/whisper-large on the 11 Sentences dataset. It achieves the following results on the evaluation set:

• Loss: 0.8472
• Wer: 169.6970

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: 4
• eval_batch_size: 8
• seed: 42
• 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: 50
• training_steps: 500
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
2.3891	8.3333	50	1.2466	21.2121
0.0553	16.6667	100	0.1580	18.1818
0.0002	25.0	150	0.1879	157.5758
0.0002	33.3333	200	0.2462	87.8788
0.0001	41.6667	250	0.3595	200.0
0.0001	50.0	300	0.5265	190.9091
0.0001	58.3333	350	0.6597	184.8485
0.0001	66.6667	400	0.7327	175.7576
0.0001	75.0	450	0.8169	172.7273
0.0001	83.3333	500	0.8472	169.6970

Framework versions

• Transformers 4.47.1
• Pytorch 2.5.1+cu124
• Datasets 3.2.0
• Tokenizers 0.21.0