vakyansh-wav2vec2-hindi-him-4200-audio-abuse-feature

This model is a fine-tuned version of Harveenchadha/vakyansh-wav2vec2-hindi-him-4200 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.6541
• Accuracy: 0.6938
• Macro F1-score: 0.6938

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: 2e-05
• train_batch_size: 16
• eval_batch_size: 16
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 64
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• num_epochs: 50

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy	Macro F1-score
6.7199	0.77	10	6.7179	0.0	0.0
6.699	1.54	20	6.6606	0.1192	0.0134
6.6237	2.31	30	6.5410	0.4499	0.0852
6.4723	3.08	40	6.3277	0.5014	0.2226
6.2904	3.85	50	6.0526	0.5041	0.3351
6.0074	4.62	60	5.7219	0.5041	0.3351
5.7233	5.38	70	5.3914	0.5041	0.3351
5.4426	6.15	80	5.0902	0.5041	0.3351
5.1776	6.92	90	4.8584	0.5041	0.3351
5.0097	7.69	100	4.6328	0.5041	0.3351
4.7851	8.46	110	4.4098	0.5041	0.3351
4.6801	9.23	120	4.2064	0.5041	0.3351
4.4144	10.0	130	3.9980	0.5041	0.3351
4.1631	10.77	140	3.7914	0.5041	0.3351
4.0093	11.54	150	3.5793	0.5041	0.3351
3.7803	12.31	160	3.3708	0.5041	0.3351
3.645	13.08	170	3.1635	0.5041	0.3351
3.3334	13.85	180	2.9564	0.5041	0.3351
3.0942	14.62	190	2.7570	0.5041	0.3351
3.0844	15.38	200	2.5619	0.5041	0.3351
2.749	16.15	210	2.3733	0.5041	0.3351
2.5448	16.92	220	2.1896	0.5041	0.3351
2.3636	17.69	230	2.0160	0.5041	0.3351
2.1303	18.46	240	1.8579	0.5041	0.3351
2.1702	19.23	250	1.7136	0.5041	0.3351
1.8911	20.0	260	1.5809	0.5041	0.3351
1.7695	20.77	270	1.4511	0.5041	0.3351
1.5466	21.54	280	1.3433	0.5041	0.3351
1.4228	22.31	290	1.2479	0.5041	0.3351
1.4089	23.08	300	1.1632	0.5041	0.3351
1.2252	23.85	310	1.0900	0.5041	0.3351
1.2236	24.62	320	1.0268	0.5041	0.3351
1.0727	25.38	330	0.9742	0.5041	0.3351
1.0036	26.15	340	0.9273	0.5041	0.3351
0.95	26.92	350	0.8892	0.5041	0.3351
0.9304	27.69	360	0.8592	0.5041	0.3351
0.9426	28.46	370	0.8355	0.5041	0.3351
0.8967	29.23	380	0.8136	0.5041	0.3351
0.862	30.0	390	0.7942	0.5041	0.3351
0.8609	30.77	400	0.7799	0.5041	0.3351
0.8013	31.54	410	0.7667	0.5041	0.3351
0.7845	32.31	420	0.7572	0.5041	0.3351
0.77	33.08	430	0.7425	0.5041	0.3351
0.7952	33.85	440	0.7309	0.5041	0.3351
0.7433	34.62	450	0.7119	0.7046	0.7036
0.7791	35.38	460	0.7002	0.7019	0.6992
0.7409	36.15	470	0.6932	0.7073	0.7029
0.7233	36.92	480	0.6887	0.6911	0.6904
0.716	37.69	490	0.6820	0.6992	0.6986
0.6994	38.46	500	0.6821	0.6883	0.6881
0.6899	39.23	510	0.6701	0.6938	0.6930
0.7014	40.0	520	0.6683	0.6965	0.6965
0.7252	40.77	530	0.6632	0.7019	0.7011
0.6763	41.54	540	0.6650	0.6883	0.6882
0.7115	42.31	550	0.6615	0.6829	0.6829
0.6703	43.08	560	0.6628	0.6911	0.6908
0.6757	43.85	570	0.6626	0.6938	0.6935
0.6672	44.62	580	0.6545	0.7046	0.7043
0.6518	45.38	590	0.6559	0.6965	0.6965
0.6661	46.15	600	0.6610	0.6802	0.6800
0.679	46.92	610	0.6580	0.6856	0.6856
0.7079	47.69	620	0.6558	0.6911	0.6911
0.7103	48.46	630	0.6536	0.6992	0.6992
0.6613	49.23	640	0.6536	0.6965	0.6965
0.6326	50.0	650	0.6541	0.6938	0.6938

Framework versions

• Transformers 4.33.0
• Pytorch 2.0.0
• Datasets 2.1.0
• Tokenizers 0.13.3