Teamim-AllNusah-whisper-small_Warmup_steps-1000_LR-1e-05_Random-True

cantillation/Teamim-AllNusah-whisper-small_Warmup_steps-1000_LR-1e-05_Random-True

Teamim-AllNusah-whisper-small_Warmup_steps-1000_LR-1e-05_Random-True is a automatic speech recognition model from cantillation. Use it when you need speech turned into text. It is set up for transformers. The card lists the license as apache-2.0.

should probably proofread and complete it, then remove this comment. --

Downloads · 30 days

20% of all-time downloads

All-time downloads

Public

Parameters

242M

9.7 GB on disk

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.safetensors967 MB · 100%

At a glance

Task: Automatic Speech Recognition
Library: transformers
License: apache-2.0
Model type: whisper
Access: Public
Created: Mar 19, 2024
Updated: Mar 20, 2024
SHA: b782bdc8

Base models

openai/whisper-small

Model card

From the Hugging Face model README

he-cantillation

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

Loss: 0.1579
Wer: 12.5795
Avg Precision Exact: 0.8918
Avg Recall Exact: 0.8920
Avg F1 Exact: 0.8915
Avg Precision Letter Shift: 0.9121
Avg Recall Letter Shift: 0.9126
Avg F1 Letter Shift: 0.9120
Avg Precision Word Level: 0.9146
Avg Recall Word Level: 0.9150
Avg F1 Word Level: 0.9144
Avg Precision Word Shift: 0.9737
Avg Recall Word Shift: 0.9750
Avg F1 Word Shift: 0.9739
Precision Median Exact: 1.0
Recall Median Exact: 1.0
F1 Median Exact: 1.0
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: 0.1
F1 Min Word Shift: 0.1176

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: 1000
training_steps: 50000
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
0.0443	0.4	5000	0.1428	18.7251	0.8438	0.8491	0.8459	0.8710	0.8764	0.8731	0.8744	0.8797	0.8765	0.9469	0.9529	0.9492	0.9167	0.9231	0.9231	1.0	1.0	1.0	0.0	0.0	0.0
0.0165	0.8	10000	0.1359	15.5691	0.8644	0.8668	0.8652	0.8882	0.8907	0.8890	0.8912	0.8938	0.8920	0.9621	0.9653	0.9632	0.9310	0.9375	0.9474	1.0	1.0	1.0	0.1429	0.1	0.1176
0.0282	1.2	15000	0.1395	14.6859	0.8710	0.8721	0.8712	0.8930	0.8941	0.8931	0.8960	0.8971	0.8961	0.9660	0.9672	0.9662	0.9375	0.9412	0.9524	1.0	1.0	1.0	0.1111	0.1	0.1176
0.0076	1.6	20000	0.1468	13.9542	0.8822	0.8850	0.8832	0.9040	0.9069	0.9051	0.9067	0.9098	0.9078	0.9672	0.9717	0.9689	0.9375	1.0	0.9565	1.0	1.0	1.0	0.1429	0.1	0.1176
0.0075	2.0	25000	0.1491	13.6179	0.8854	0.8860	0.8853	0.9067	0.9073	0.9066	0.9092	0.9101	0.9092	0.9685	0.9698	0.9686	1.0	1.0	0.9630	1.0	1.0	1.0	0.1429	0.1	0.1176
0.0033	2.4	30000	0.1531	13.2114	0.8901	0.8900	0.8897	0.9102	0.9101	0.9097	0.9126	0.9127	0.9122	0.9712	0.9723	0.9713	1.0	1.0	1.0	1.0	1.0	1.0	0.1429	0.1	0.1176
0.0045	2.8	35000	0.1535	13.0488	0.8920	0.8921	0.8916	0.9129	0.9131	0.9126	0.9153	0.9155	0.9150	0.9726	0.9734	0.9725	1.0	1.0	0.9697	1.0	1.0	1.0	0.1429	0.1	0.1176
0.0011	3.2	40000	0.1543	12.9084	0.8924	0.8940	0.8928	0.9131	0.9149	0.9136	0.9158	0.9174	0.9162	0.9713	0.9735	0.9719	1.0	1.0	1.0	1.0	1.0	1.0	0.0833	0.0909	0.0870
0.0007	3.6	45000	0.1571	12.7679	0.8907	0.8926	0.8913	0.9118	0.9140	0.9125	0.9146	0.9165	0.9152	0.9724	0.9750	0.9733	1.0	1.0	1.0	1.0	1.0	1.0	0.1429	0.1	0.1176
0.0005	4.0	50000	0.1579	12.5795	0.8918	0.8920	0.8915	0.9121	0.9126	0.9120	0.9146	0.9150	0.9144	0.9737	0.9750	0.9739	1.0	1.0	1.0	1.0	1.0	1.0	0.1429	0.1	0.1176

Framework versions

Transformers 4.39.0.dev0
Pytorch 2.2.1+cu121
Datasets 2.16.1
Tokenizers 0.15.0

Task: Automatic Speech Recognition
Library: transformers
Type: whisper
License: apache-2.0
Languages: he
Created: Mar 19, 2024
Updated: Mar 20, 2024