videomae-base-finetuned-isl-numbers

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: 1.1287
• Accuracy: 0.6444

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: 8
• eval_batch_size: 8
• 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: 1100

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.4795	0.02	22	2.4767	0.0256
2.4249	1.02	44	2.4351	0.1026
2.4561	2.02	66	2.4196	0.1026
2.3841	3.02	88	2.3735	0.1026
2.5186	4.02	110	2.4258	0.0769
2.3806	5.02	132	2.3214	0.1538
2.3579	6.02	154	2.2858	0.1538
2.2955	7.02	176	2.1729	0.1795
2.1351	8.02	198	1.9503	0.3333
2.1626	9.02	220	2.1922	0.2051
2.0905	10.02	242	1.8453	0.3333
1.7091	11.02	264	1.6305	0.4872
1.6316	12.02	286	1.6529	0.3333
1.6399	13.02	308	1.7789	0.2308
1.5139	14.02	330	1.6245	0.3590
1.3315	15.02	352	1.6540	0.2821
1.0726	16.02	374	1.7507	0.2821
1.1432	17.02	396	1.6282	0.3333
1.144	18.02	418	1.3435	0.5128
0.987	19.02	440	0.8631	0.7949
0.8152	20.02	462	1.0812	0.5897
0.8175	21.02	484	1.4527	0.4359
0.7587	22.02	506	1.2309	0.5128
0.6255	23.02	528	1.1940	0.4872
0.6867	24.02	550	0.9270	0.5385
0.7537	25.02	572	0.6586	0.7436
0.6147	26.02	594	0.7935	0.7179
0.4602	27.02	616	0.9698	0.6154
0.482	28.02	638	0.9328	0.6410
0.3436	29.02	660	0.9947	0.6154
0.336	30.02	682	0.8127	0.6410
0.3952	31.02	704	0.5542	0.8205
0.2922	32.02	726	1.3266	0.5897
0.2998	33.02	748	0.9621	0.6410
0.2824	34.02	770	0.7805	0.7436
0.2971	35.02	792	0.4700	0.8462
0.1746	36.02	814	0.6059	0.8205
0.1325	37.02	836	0.4568	0.7436
0.2452	38.02	858	0.3495	0.8462
0.161	39.02	880	0.2546	0.9231
0.1788	40.02	902	0.3275	0.8974
0.201	41.02	924	0.3987	0.8205
0.259	42.02	946	0.5395	0.7692
0.112	43.02	968	0.4591	0.8462
0.0622	44.02	990	0.3455	0.8462
0.1307	45.02	1012	0.5513	0.7436
0.0924	46.02	1034	0.6709	0.7436
0.056	47.02	1056	0.4471	0.8205
0.089	48.02	1078	0.3860	0.8205
0.1798	49.02	1100	0.4313	0.8462

Framework versions

• Transformers 4.40.0
• Pytorch 2.1.0+cu121
• Datasets 2.18.0
• Tokenizers 0.19.1