vakyansh-wav2vec2-odia-orm-100-audio-abuse-feature

This model is a fine-tuned version of Harveenchadha/vakyansh-wav2vec2-odia-orm-100 on the None dataset. It achieves the following results on the evaluation set:

• Loss: 0.7299
• Accuracy: 0.7014
• Macro F1-score: 0.6792

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
6.7078	0.78	10	6.6948	0.0	0.0
6.6539	1.57	20	6.5580	0.2	0.0342
6.5111	2.35	30	6.3377	0.5726	0.3641
6.268	3.14	40	6.0361	0.5726	0.3641
6.0748	3.92	50	5.7417	0.5726	0.3641
5.8205	4.71	60	5.4985	0.5726	0.3641
5.6051	5.49	70	5.2743	0.5726	0.3641
5.3589	6.27	80	5.0823	0.5726	0.3641
5.2019	7.06	90	4.8953	0.5726	0.3641
5.0528	7.84	100	4.7077	0.5726	0.3641
4.868	8.63	110	4.5244	0.5726	0.3641
4.7081	9.41	120	4.3347	0.5726	0.3641
4.437	10.2	130	4.1455	0.5726	0.3641
4.3225	10.98	140	3.9551	0.5726	0.3641
4.0945	11.76	150	3.7694	0.5726	0.3641
4.014	12.55	160	3.5710	0.5726	0.3641
3.8491	13.33	170	3.3814	0.5726	0.3641
3.4724	14.12	180	3.1873	0.5726	0.3641
3.2728	14.9	190	2.9999	0.5726	0.3641
3.1948	15.69	200	2.8224	0.5726	0.3641
2.9968	16.47	210	2.6368	0.5726	0.3641
2.6739	17.25	220	2.4462	0.5726	0.3641
2.561	18.04	230	2.2871	0.5726	0.3641
2.5101	18.82	240	2.1260	0.5726	0.3641
2.3307	19.61	250	1.9620	0.5726	0.3641
2.1022	20.39	260	1.8260	0.5726	0.3641
1.9909	21.18	270	1.6933	0.5726	0.3641
1.766	21.96	280	1.5644	0.5726	0.3641
1.7143	22.75	290	1.4669	0.5726	0.3641
1.5073	23.53	300	1.3482	0.5726	0.3641
1.6055	24.31	310	1.2643	0.5726	0.3641
1.321	25.1	320	1.1930	0.5726	0.3641
1.2165	25.88	330	1.1128	0.5726	0.3641
1.1484	26.67	340	1.0493	0.6712	0.6033
1.1413	27.45	350	0.9925	0.7096	0.6737
1.0462	28.24	360	0.9471	0.6877	0.6190
0.9667	29.02	370	0.9209	0.7123	0.6869
0.9918	29.8	380	0.8892	0.7205	0.6953
0.9112	30.59	390	0.8414	0.7123	0.6705
0.8666	31.37	400	0.8291	0.7123	0.6836
0.8096	32.16	410	0.8284	0.6959	0.6501
0.7987	32.94	420	0.7729	0.7425	0.7270
0.7529	33.73	430	0.7542	0.7260	0.7023
0.7605	34.51	440	0.7535	0.7260	0.7043
0.7011	35.29	450	0.7882	0.6959	0.6891
0.6868	36.08	460	0.7378	0.7260	0.7013
0.6858	36.86	470	0.7518	0.7096	0.6865
0.7546	37.65	480	0.7163	0.7342	0.7108
0.6717	38.43	490	0.7158	0.7397	0.7158
0.7048	39.22	500	0.7755	0.6575	0.6487
0.6767	40.0	510	0.7469	0.7068	0.6798
0.6621	40.78	520	0.7166	0.7205	0.7020
0.6639	41.57	530	0.7143	0.7151	0.6934
0.5988	42.35	540	0.7547	0.6767	0.6661
0.6179	43.14	550	0.7394	0.7014	0.6820
0.7033	43.92	560	0.7312	0.6986	0.6757
0.6076	44.71	570	0.7331	0.6904	0.6674
0.602	45.49	580	0.7341	0.6932	0.6718
0.545	46.27	590	0.7363	0.6932	0.6738
0.5881	47.06	600	0.7299	0.7014	0.6792

Framework versions

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