whisper-base-aug-24-mar-v2

This model is a fine-tuned version of PhanithLIM/whisper-base-aug-22-mar-v1 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 0.1307
• Wer: 93.2722

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: 16
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: constant
• lr_scheduler_warmup_steps: 1000
• num_epochs: 10

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
0.2666	1.0	56	0.1865	103.5168
0.1775	2.0	112	0.1622	98.3945
0.1411	3.0	168	0.1467	98.5474
0.1174	4.0	224	0.1420	94.8012
0.0989	5.0	280	0.1362	95.1070
0.085	6.0	336	0.1352	100.0
0.0739	7.0	392	0.1328	92.4312
0.0653	8.0	448	0.1340	96.1774
0.0576	9.0	504	0.1320	94.9541
0.051	9.8288	550	0.1307	93.2722

Framework versions

• Transformers 4.50.0
• Pytorch 2.4.1+cu121
• Datasets 3.4.1
• Tokenizers 0.21.1