videomae-base-finetuned-scratch_1

This model is a fine-tuned version of MCG-NJU/videomae-base on an unknown dataset. It achieves the following results on the evaluation set:

• Loss: 1.6254
• Accuracy: 0.7624

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: 5e-05
• train_batch_size: 12
• eval_batch_size: 12
• seed: 42
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• training_steps: 30192

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.5872	0.01	420	0.6884	0.5518
0.5358	1.01	840	0.6766	0.6194
0.5339	2.01	1260	0.6295	0.6430
0.4483	3.01	1680	0.7323	0.5957
0.4654	4.01	2100	0.7020	0.6486
0.3897	5.01	2520	0.7636	0.6498
0.3386	6.01	2940	0.8877	0.6610
0.3601	7.01	3360	0.8791	0.6486
0.3401	8.01	3780	0.7403	0.6633
0.3113	9.01	4200	0.7316	0.6959
0.2096	10.01	4620	0.9519	0.6982
0.1537	11.01	5040	0.9116	0.7016
0.1113	12.01	5460	1.0047	0.6971
0.3247	13.01	5880	1.2167	0.6847
0.171	14.01	6300	0.9337	0.7027
0.3076	15.01	6720	1.1811	0.7207
0.2927	16.01	7140	1.0953	0.7218
0.1679	17.01	7560	1.2948	0.7207
0.1523	18.01	7980	1.3632	0.7016
0.1059	19.01	8400	1.2915	0.7185
0.1741	20.01	8820	1.2315	0.7432
0.0629	21.01	9240	1.3948	0.7230
0.0075	22.01	9660	1.1435	0.7376
0.1692	23.01	10080	1.3998	0.7128
0.0347	24.01	10500	1.4803	0.7027
0.0396	25.01	10920	1.6457	0.7005
0.0074	26.01	11340	1.5602	0.7050
0.1256	27.01	11760	1.3965	0.7173
0.0021	28.01	12180	1.4514	0.7342
0.0476	29.01	12600	1.2915	0.7173
0.0065	30.01	13020	1.3397	0.7095
0.0435	31.01	13440	1.8912	0.6948
0.0268	32.01	13860	1.5767	0.7286
0.0487	33.01	14280	1.6439	0.6948
0.0448	34.01	14700	1.5990	0.7354
0.0166	35.01	15120	1.3866	0.7466
0.1029	36.01	15540	1.7427	0.7106
0.0678	37.01	15960	1.4194	0.7365
0.0007	38.01	16380	1.9137	0.7072
0.0602	39.01	16800	1.6180	0.7309
0.0977	40.01	17220	1.5710	0.7354
0.0606	41.01	17640	1.3908	0.7342
0.1046	42.01	18060	1.7846	0.7252
0.0004	43.01	18480	1.6396	0.7241
0.0881	44.01	18900	1.6206	0.7196
0.0934	45.01	19320	1.6994	0.7320
0.0001	46.01	19740	2.0068	0.7162
0.024	47.01	20160	1.5350	0.7376
0.017	48.01	20580	1.8864	0.7162
0.121	49.01	21000	1.6862	0.7230
0.0001	50.01	21420	1.8462	0.7365
0.0319	51.01	21840	1.9072	0.7286
0.065	52.01	22260	1.6631	0.7556
0.0424	53.01	22680	1.9177	0.7399
0.0001	54.01	23100	1.8990	0.7365
0.0	55.01	23520	2.0622	0.7455
0.0582	56.01	23940	1.4821	0.7444
0.0001	57.01	24360	1.6254	0.7624
0.0	58.01	24780	1.8024	0.7545
0.0486	59.01	25200	1.6804	0.7545
0.0001	60.01	25620	1.7991	0.7523
0.0	61.01	26040	1.8281	0.7511
0.0022	62.01	26460	1.8172	0.75
0.0001	63.01	26880	1.9532	0.7489
0.0	64.01	27300	1.9209	0.7477
0.0	65.01	27720	1.9100	0.7579
0.0	66.01	28140	1.9572	0.7534
0.0007	67.01	28560	2.0380	0.75
0.0627	68.01	28980	1.8911	0.7579
0.0002	69.01	29400	1.9255	0.75
0.0	70.01	29820	1.9195	0.7568
0.0	71.01	30192	1.9317	0.75

Framework versions

• Transformers 4.39.0
• Pytorch 2.2.1+cu121
• Datasets 2.18.0
• Tokenizers 0.15.2