cantillation/Teamim-small_WeightDecay-0.05_Augmented_New-Data_nusach-yerushalmi_date-24-07-2024

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: 0.2759
• Wer: 12.0860
• Avg Precision Exact: 0.9045
• Avg Recall Exact: 0.9054
• Avg F1 Exact: 0.9045
• Avg Precision Letter Shift: 0.9223
• Avg Recall Letter Shift: 0.9233
• Avg F1 Letter Shift: 0.9224
• Avg Precision Word Level: 0.9250
• Avg Recall Word Level: 0.9259
• Avg F1 Word Level: 0.9250
• Avg Precision Word Shift: 0.9777
• Avg Recall Word Shift: 0.9785
• Avg F1 Word Shift: 0.9777
• 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: 0.0
• 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-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: 500
• training_steps: 200000
• 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	0.0004	1	6.8177	106.5214	0.0004	0.0012	0.0006	0.0038	0.0036	0.0033	0.0030	0.0121	0.0043	0.0322	0.0342	0.0300	0.0	0.0	0.0	0.0909	0.3333	0.1429	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
0.0059	3.7023	10000	0.1748	15.8840	0.8772	0.8813	0.8786	0.9013	0.9056	0.9028	0.9063	0.9103	0.9077	0.9648	0.9693	0.9663	0.9286	0.9375	0.9474	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
0.0045	7.4047	20000	0.2047	15.1038	0.8686	0.8670	0.8673	0.8906	0.8892	0.8894	0.8952	0.8935	0.8938	0.9722	0.9711	0.9710	0.9375	0.9333	0.9524	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
0.001	11.1070	30000	0.2024	13.8083	0.8862	0.8876	0.8863	0.9076	0.9094	0.9080	0.9109	0.9127	0.9113	0.9743	0.9767	0.9749	1.0	1.0	0.9600	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
0.001	14.8093	40000	0.2188	13.8083	0.8924	0.8918	0.8916	0.9125	0.9118	0.9116	0.9166	0.9156	0.9155	0.9733	0.9730	0.9726	1.0	1.0	0.9630	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
0.0005	18.5117	50000	0.2256	13.6464	0.8921	0.8937	0.8924	0.9131	0.9148	0.9135	0.9161	0.9176	0.9164	0.9760	0.9774	0.9762	1.0	1.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
0.0012	22.2140	60000	0.2194	12.8515	0.8896	0.8917	0.8902	0.9089	0.9110	0.9095	0.9116	0.9138	0.9122	0.9748	0.9780	0.9759	1.0	1.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
0.0006	25.9163	70000	0.2265	13.0870	0.8981	0.9013	0.8992	0.9191	0.9224	0.9203	0.9219	0.9249	0.9229	0.9756	0.9776	0.9761	1.0	1.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
0.0001	29.6187	80000	0.2249	13.0870	0.8938	0.8961	0.8945	0.9139	0.9163	0.9146	0.9169	0.9191	0.9175	0.9749	0.9764	0.9752	1.0	1.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
0.0001	33.3210	90000	0.2379	13.2342	0.8960	0.8987	0.8969	0.9160	0.9189	0.9169	0.9197	0.9224	0.9206	0.9759	0.9780	0.9764	1.0	1.0	0.9697	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
0.0	37.0233	100000	0.2302	13.1312	0.8910	0.8958	0.8930	0.9121	0.9171	0.9142	0.9149	0.9195	0.9167	0.9742	0.9786	0.9759	1.0	1.0	0.9677	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
0.0004	40.7257	110000	0.2294	12.9987	0.9032	0.9028	0.9025	0.9220	0.9216	0.9213	0.9255	0.9249	0.9247	0.9762	0.9773	0.9763	1.0	1.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
0.0001	44.4280	120000	0.2322	12.6601	0.9038	0.9045	0.9037	0.9234	0.9242	0.9233	0.9262	0.9270	0.9262	0.9766	0.9784	0.9770	1.0	1.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
0.0	48.1303	130000	0.2362	12.5129	0.9054	0.9058	0.9051	0.9241	0.9247	0.9239	0.9277	0.9284	0.9276	0.9763	0.9777	0.9766	1.0	1.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
0.0	51.8327	140000	0.2430	13.1753	0.8973	0.8993	0.8978	0.9184	0.9205	0.9189	0.9216	0.9237	0.9221	0.9766	0.9783	0.9770	1.0	1.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
0.0005	55.5350	150000	0.2325	12.7926	0.9032	0.9032	0.9028	0.9226	0.9228	0.9223	0.9251	0.9252	0.9247	0.9781	0.9785	0.9779	1.0	1.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
0.0	59.2373	160000	0.2428	12.2332	0.9090	0.9104	0.9093	0.9275	0.9289	0.9278	0.9301	0.9315	0.9304	0.9773	0.9791	0.9778	1.0	1.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
0.0	62.9397	170000	0.2499	12.1301	0.9067	0.9081	0.9070	0.9246	0.9261	0.9249	0.9273	0.9286	0.9275	0.9775	0.9794	0.9780	1.0	1.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
0.0003	66.6420	180000	0.2572	12.2185	0.9050	0.9049	0.9045	0.9238	0.9238	0.9234	0.9265	0.9263	0.9260	0.9785	0.9784	0.9780	1.0	1.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
0.0	70.3443	190000	0.2704	12.1449	0.9058	0.9068	0.9059	0.9237	0.9247	0.9238	0.9263	0.9273	0.9264	0.9775	0.9787	0.9777	1.0	1.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0
0.0	74.0466	200000	0.2759	12.0860	0.9045	0.9054	0.9045	0.9223	0.9233	0.9224	0.9250	0.9259	0.9250	0.9777	0.9785	0.9777	1.0	1.0	1.0	1.0	1.0	1.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0	0.0

Framework versions

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