miosipof/asr_Synth_EasyCall_v0

Whisper Medium

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

• Loss: 0.8596
• Wer: 62.5806

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: 0.0003
• train_batch_size: 16
• eval_batch_size: 8
• seed: 42
• gradient_accumulation_steps: 2
• total_train_batch_size: 32
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_steps: 100
• num_epochs: 5
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer
5.3706	0.6897	50	4.0385	81.2903
1.6886	1.3793	100	1.1083	72.5806
0.6572	2.0690	150	0.9583	63.2258
0.4765	2.7586	200	0.9143	121.9355
0.3623	3.4483	250	0.8818	118.0645
0.2794	4.1379	300	0.8554	63.2258
0.2252	4.8276	350	0.8596	62.5806

Framework versions

• PEFT 0.14.0
• Transformers 4.45.2
• Pytorch 2.2.0
• Datasets 3.2.0
• Tokenizers 0.20.3