metadata

language:
  - he
base_model: ivrit-ai/whisper-v2-pd1-e1
tags:
  - hf-asr-leaderboard
  - generated_from_trainer
metrics:
  - wer
model-index:
  - name: he-cantillation
    results: []

he-cantillation

This model is a fine-tuned version of ivrit-ai/whisper-v2-pd1-e1 on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 0.1099
Wer: 16.2686
Avg Precision Exact: 0.8645
Avg Recall Exact: 0.8618
Avg F1 Exact: 0.8626
Avg Precision Letter Shift: 0.8838
Avg Recall Letter Shift: 0.8813
Avg F1 Letter Shift: 0.8819
Avg Precision Word Level: 0.8869
Avg Recall Word Level: 0.8850
Avg F1 Word Level: 0.8853
Avg Precision Word Shift: 0.9412
Avg Recall Word Shift: 0.9408
Avg F1 Word Shift: 0.9403
Precision Median Exact: 0.9231
Recall Median Exact: 0.9231
F1 Median Exact: 0.9286
Precision Max Exact: 1.0
Recall Max Exact: 1.0
F1 Max Exact: 1.0
Precision Min Exact: 0.0
Recall Min Exact: 0.0
F1 Min Exact: 0.0
Precision Min Letter Shift: 0.0
Recall Min Letter Shift: 0.0
F1 Min Letter Shift: 0.0
Precision Min Word Level: 0.0
Recall Min Word Level: 0.0
F1 Min Word Level: 0.0
Precision Min Word Shift: 0.1429
Recall Min Word Shift: 0.0909
F1 Min Word Shift: 0.1111

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: 8
eval_batch_size: 32
seed: 42
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 10
training_steps: 3000
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Wer	Avg Precision Exact	Avg Recall Exact	Avg F1 Exact	Avg Precision Letter Shift	Avg Recall Letter Shift	Avg F1 Letter Shift	Avg Precision Word Level	Avg Recall Word Level	Avg F1 Word Level	Avg Precision Word Shift	Avg Recall Word Shift	Avg F1 Word Shift	Precision Median Exact	Recall Median Exact	F1 Median Exact	Precision Max Exact	Recall Max Exact	F1 Max Exact	Precision Min Word Shift	Recall Min Word Shift	F1 Min Word Shift
0.1552	0.0517	1000	0.1843	26.9266	0.7790	0.7668	0.7718	0.8099	0.7974	0.8025	0.8153	0.8044	0.8088	0.8935	0.8895	0.8902	0.8462	0.8462	0.8462	1.0	1.0	1.0	0.1429	0.0909	0.1111
0.0973	0.1033	2000	0.1317	19.9251	0.8323	0.8338	0.8323	0.8540	0.8558	0.8542	0.8583	0.8605	0.8587	0.9204	0.9258	0.9223	0.9091	0.9091	0.9091	1.0	1.0	1.0	0.0	0.0	0.0
0.0625	0.1550	3000	0.1099	16.2686	0.8645	0.8618	0.8626	0.8838	0.8813	0.8819	0.8869	0.8850	0.8853	0.9412	0.9408	0.9403	0.9231	0.9231	0.9286	1.0	1.0	1.0	0.1429	0.0909	0.1111

Framework versions

Transformers 4.41.2
Pytorch 2.2.1
Datasets 2.20.0
Tokenizers 0.19.1