You can deploy this model and start earning money today!

callmesan/
vakyansh-wav2vec2-bengali-bnm-200-audio-abuse-feature

Name: vakyansh-wav2vec2-bengali-bnm-200-audio-abuse-feature
Availability: InStock

generated_from_trainer

Harveenchadha/vakyansh-wav2vec2-bengali-bnm-200

finetune

endpoints_compatible

Views

No views yet

vakyansh-wav2vec2-bengali-bnm-200-audio-abuse-feature

This model is a fine-tuned version of Harveenchadha/vakyansh-wav2vec2-bengali-bnm-200 on the None dataset. It achieves the following results on the evaluation set:

Loss: 0.8024
Accuracy: 0.6459
Macro F1-score: 0.6339

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
6.7346	0.77	10	6.7291	0.0	0.0
6.6926	1.54	20	6.6304	0.0027	0.0004
6.5587	2.31	30	6.4243	0.5730	0.2830
6.3449	3.08	40	6.1051	0.5649	0.5571
6.1232	3.85	50	5.7862	0.4216	0.3430
5.8191	4.62	60	5.5131	0.4027	0.2908
5.592	5.38	70	5.2719	0.5216	0.5022
5.3414	6.15	80	5.0558	0.6189	0.6186
5.1331	6.92	90	4.8552	0.6865	0.6852
4.98	7.69	100	4.6603	0.6568	0.6567
4.7844	8.46	110	4.4634	0.6703	0.6702
4.7028	9.23	120	4.2715	0.6568	0.6567
4.4476	10.0	130	4.0733	0.6297	0.6280
4.2098	10.77	140	3.8749	0.6108	0.6041
4.0715	11.54	150	3.6803	0.5027	0.4564
3.8545	12.31	160	3.4603	0.6649	0.6648
3.708	13.08	170	3.2559	0.6541	0.6534
3.4318	13.85	180	3.0493	0.6676	0.6675
3.1874	14.62	190	2.8456	0.6838	0.6837
3.1887	15.38	200	2.6625	0.5595	0.5384
2.8359	16.15	210	2.4679	0.5757	0.5604
2.6265	16.92	220	2.2662	0.6892	0.6841
2.4536	17.69	230	2.0843	0.6649	0.6644
2.2288	18.46	240	1.9218	0.6459	0.6431
2.2955	19.23	250	1.7633	0.6595	0.6578
1.9739	20.0	260	1.6105	0.6730	0.6671
1.8575	20.77	270	1.4855	0.6378	0.6351
1.607	21.54	280	1.3582	0.6649	0.6646
1.4831	22.31	290	1.2425	0.6676	0.6646
1.4484	23.08	300	1.1522	0.6703	0.6660
1.2517	23.85	310	1.0688	0.6595	0.6554
1.2793	24.62	320	1.0006	0.6541	0.6523
1.0722	25.38	330	0.9486	0.6568	0.6543
0.9888	26.15	340	0.9292	0.6135	0.6135
0.9134	26.92	350	0.8580	0.6514	0.6492
0.9208	27.69	360	0.8352	0.6649	0.6646
0.966	28.46	370	0.8220	0.6162	0.6160
0.8746	29.23	380	0.8064	0.6568	0.6420
0.8619	30.0	390	0.7856	0.6405	0.5942
0.841	30.77	400	0.7612	0.6459	0.6020
0.7629	31.54	410	0.7441	0.6459	0.6434
0.6736	32.31	420	0.7610	0.6568	0.6562
0.6579	33.08	430	0.7624	0.6514	0.6456
0.7514	33.85	440	0.7374	0.6649	0.6467
0.6579	34.62	450	0.7503	0.6541	0.6471
0.6864	35.38	460	0.8286	0.5892	0.5889
0.6863	36.15	470	0.7393	0.6541	0.6396
0.6224	36.92	480	0.7427	0.6541	0.6507
0.6255	37.69	490	0.7495	0.6405	0.6268
0.5295	38.46	500	0.7787	0.6486	0.6385
0.5549	39.23	510	0.7909	0.6378	0.6360
0.5752	40.0	520	0.7631	0.6459	0.6361
0.616	40.77	530	0.7636	0.6432	0.6390
0.5038	41.54	540	0.7847	0.6514	0.6372
0.5935	42.31	550	0.7837	0.6595	0.6461
0.5453	43.08	560	0.7804	0.6405	0.6330
0.5378	43.85	570	0.7928	0.6514	0.6338
0.4852	44.62	580	0.8249	0.6324	0.6285
0.5198	45.38	590	0.8065	0.6459	0.6186
0.5067	46.15	600	0.8210	0.6162	0.6107
0.5533	46.92	610	0.8053	0.6432	0.6300
0.6282	47.69	620	0.7970	0.6459	0.6316
0.5617	48.46	630	0.8095	0.6243	0.6165
0.5016	49.23	640	0.8038	0.6378	0.6274
0.467	50.0	650	0.8024	0.6459	0.6339

Framework versions

Transformers 4.33.0
Pytorch 2.0.0
Datasets 2.1.0
Tokenizers 0.13.3

callmesan/vakyansh-wav2vec2-bengali-bnm-200-audio-abuse-feature

vakyansh-wav2vec2-bengali-bnm-200-audio-abuse-feature

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

callmesan/
vakyansh-wav2vec2-bengali-bnm-200-audio-abuse-feature