cantillation/Teamim-tiny_DropOut-0.5_Augmented_Combined-Data_date-07-07-2024_08-34

he-cantillation

This model is a fine-tuned version of cantillation/Teamim-tiny_DropOut-0.5_Augmented_Combined-Data_date-06-07-2024_20-19 on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 3.9592
• Wer: 100.0
• Avg Precision Exact: 0.0053
• Avg Recall Exact: 0.0482
• Avg F1 Exact: 0.0094
• Avg Precision Letter Shift: 0.0129
• Avg Recall Letter Shift: 0.1183
• Avg F1 Letter Shift: 0.0228
• Avg Precision Word Level: 0.0173
• Avg Recall Word Level: 0.1573
• Avg F1 Word Level: 0.0306
• Avg Precision Word Shift: 0.0334
• Avg Recall Word Shift: 0.2899
• Avg F1 Word Shift: 0.0588
• Precision Median Exact: 0.0
• Recall Median Exact: 0.0
• F1 Median Exact: 0.0
• Precision Max Exact: 0.1667
• Recall Max Exact: 1.0
• F1 Max Exact: 0.25
• 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: 0.0
• Recall Min Word Shift: 0.0
• F1 Min Word Shift: 0.0

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-06
• 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 Exact	Recall Min Exact	F1 Min Exact	Precision Min Letter Shift	Recall Min Letter Shift	F1 Min Letter Shift	Precision Min Word Level	Recall Min Word Level	F1 Min Word Level	Precision Min Word Shift	Recall Min Word Shift	F1 Min Word Shift
No log	0.0001	1	0.2806	16.9452	0.8385	0.8403	0.8390	0.8609	0.8628	0.8613	0.8657	0.8679	0.8663	0.9468	0.9505	0.9481	0.9231	0.9231	0.9286	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0833	0.0769	0.08
6.379	0.1033	2000	1.1953	120.7433	0.1765	0.2759	0.1896	0.2338	0.3541	0.2490	0.2533	0.3945	0.2751	0.4039	0.6029	0.4412	0.1538	0.2	0.1667	1.0	1.0	0.9412	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
5.4428	0.2067	4000	4.7348	100.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
5.1556	0.3100	6000	4.6051	100.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
4.935	0.4134	8000	4.4172	100.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0	0	0	0	0	0	0	0	0	0	0	0	0	0	0
4.7728	0.5167	10000	4.1229	100.0	0.0	0.0	0.0	0.0003	0.0026	0.0005	0.0008	0.0092	0.0015	0.0012	0.0141	0.0022	0.0	0.0	0.0	0	0	0	0	0	0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
4.636	0.6200	12000	4.0112	100.0378	0.0071	0.0423	0.0116	0.0155	0.0901	0.0251	0.0260	0.1528	0.0422	0.0567	0.3257	0.0922	0.0	0.0	0.0	0.2222	1.0	0.3333	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
4.4949	0.7234	14000	4.0080	100.0	0.0040	0.0390	0.0072	0.0097	0.0899	0.0171	0.0180	0.1677	0.0317	0.0349	0.3220	0.0615	0.0	0.0	0.0	0.1818	1.0	0.3077	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
4.4359	0.8267	16000	4.0252	100.0	0.0031	0.0347	0.0057	0.0075	0.0849	0.0138	0.0144	0.1589	0.0262	0.0250	0.2827	0.0457	0.0	0.0	0.0	0.1429	1.0	0.25	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
4.3834	0.9300	18000	4.0642	100.0	0.0025	0.0273	0.0045	0.0063	0.0700	0.0115	0.0125	0.1372	0.0228	0.0225	0.2534	0.0412	0.0	0.0	0.0	0.1429	1.0	0.25	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
4.317	1.0334	20000	4.0775	100.0	0.0017	0.0183	0.0031	0.0043	0.0470	0.0079	0.0080	0.0860	0.0146	0.0150	0.1663	0.0274	0.0	0.0	0.0	0.1429	1.0	0.25	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
4.269	1.1367	22000	4.1070	100.0	0.0010	0.0108	0.0018	0.0025	0.0266	0.0046	0.0044	0.0447	0.0079	0.0080	0.0834	0.0144	0.0	0.0	0.0	0.125	1.0	0.2222	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
4.2529	1.2401	24000	4.0914	100.0	0.0008	0.0086	0.0015	0.0021	0.0215	0.0038	0.0033	0.0339	0.0060	0.0064	0.0658	0.0115	0.0	0.0	0.0	0.125	1.0	0.2222	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
4.1986	1.3434	26000	4.0861	100.0	0.0012	0.0131	0.0022	0.0025	0.0255	0.0044	0.0038	0.0404	0.0070	0.0081	0.0865	0.0146	0.0	0.0	0.0	0.1429	1.0	0.25	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
4.1805	1.4467	28000	4.0716	100.0	0.0016	0.0171	0.0029	0.0034	0.0362	0.0062	0.0055	0.0579	0.0099	0.0103	0.1118	0.0188	0.0	0.0	0.0	0.1429	1.0	0.25	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
4.1489	1.5501	30000	4.0526	100.0	0.0027	0.0306	0.0050	0.0065	0.0701	0.0117	0.0096	0.1033	0.0174	0.0174	0.1901	0.0316	0.0	0.0	0.0	0.1429	1.0	0.25	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
4.0979	1.6534	32000	4.0399	100.0	0.0034	0.0389	0.0062	0.0082	0.0924	0.0149	0.0121	0.1331	0.0220	0.0217	0.2389	0.0394	0.0	0.0	0.0	0.2	1.0	0.3333	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
4.0667	1.7567	34000	4.0224	100.0	0.0042	0.0473	0.0076	0.0099	0.1091	0.0179	0.0143	0.1549	0.0259	0.0258	0.2749	0.0465	0.0	0.0	0.0	0.2	1.0	0.3333	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
4.0602	1.8601	36000	4.0042	100.0	0.0040	0.0466	0.0073	0.0103	0.1182	0.0187	0.0142	0.1606	0.0259	0.0254	0.2818	0.0461	0.0	0.0	0.0	0.2	1.0	0.3333	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
4.0584	1.9634	38000	3.9922	100.0	0.0038	0.0441	0.0069	0.0104	0.1215	0.0190	0.0140	0.1623	0.0256	0.0244	0.2807	0.0446	0.0	0.0	0.0	0.2	1.0	0.3333	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
4.0015	2.0668	40000	3.9825	100.0	0.0042	0.0506	0.0077	0.0111	0.1293	0.0203	0.0149	0.1726	0.0272	0.0254	0.2923	0.0464	0.0	0.0	0.0	0.1429	1.0	0.25	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
4.0073	2.1701	42000	3.9722	100.0	0.0044	0.0508	0.0080	0.0112	0.1285	0.0205	0.0149	0.1704	0.0272	0.0255	0.2891	0.0465	0.0	0.0	0.0	0.1429	1.0	0.25	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
4.0284	2.2734	44000	3.9651	100.0	0.0047	0.0506	0.0084	0.0116	0.1258	0.0211	0.0156	0.1669	0.0281	0.0278	0.2904	0.0501	0.0	0.0	0.0	0.1429	1.0	0.25	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
3.9935	2.3768	46000	3.9619	100.0	0.0051	0.0507	0.0092	0.0122	0.1230	0.0219	0.0165	0.1640	0.0295	0.0298	0.2889	0.0532	0.0	0.0	0.0	0.1667	1.0	0.25	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
4.0324	2.4801	48000	3.9600	100.0	0.0052	0.0485	0.0092	0.0126	0.1197	0.0225	0.0171	0.1597	0.0304	0.0326	0.2909	0.0577	0.0	0.0	0.0	0.1667	1.0	0.25	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
4.003	2.5834	50000	3.9592	100.0	0.0053	0.0482	0.0094	0.0129	0.1183	0.0228	0.0173	0.1573	0.0306	0.0334	0.2899	0.0588	0.0	0.0	0.0	0.1667	1.0	0.25	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0

Framework versions

• Transformers 4.41.2
• Pytorch 2.2.1
• Datasets 2.20.0
• Tokenizers 0.19.1