MAE-CT-M1N0-M12_v8_split1

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

• Loss: 1.4264
• Accuracy: 0.7297

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: 1e-05
• train_batch_size: 4
• eval_batch_size: 4
• seed: 42
• optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
• lr_scheduler_type: linear
• lr_scheduler_warmup_ratio: 0.1
• training_steps: 6400

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
0.6776	0.0102	65	0.6899	0.5652
0.672	1.0102	130	0.7048	0.5652
0.6127	2.0102	195	0.7328	0.5652
0.6091	3.0102	260	0.6351	0.5652
0.4716	4.0102	325	0.6692	0.5652
0.3767	5.0102	390	1.0344	0.5652
0.444	6.0102	455	0.5516	0.6522
0.5153	7.0102	520	0.5306	0.8261
0.5645	8.0102	585	0.8665	0.5652
0.5957	9.0102	650	1.0774	0.5652
0.265	10.0102	715	1.1522	0.6522
0.2538	11.0102	780	1.1066	0.6087
0.3453	12.0102	845	1.3690	0.6087
0.3042	13.0102	910	0.3015	0.9130
0.1095	14.0102	975	1.0894	0.7391
0.3343	15.0102	1040	1.0325	0.7391
0.4758	16.0102	1105	0.2653	0.9130
0.2326	17.0102	1170	0.3844	0.9130
0.1366	18.0102	1235	0.3475	0.9565
0.0036	19.0102	1300	0.6067	0.9130
0.0005	20.0102	1365	1.2899	0.7826
0.0364	21.0102	1430	2.7527	0.5652
0.1536	22.0102	1495	1.9413	0.6522
0.03	23.0102	1560	2.2468	0.6522
0.0004	24.0102	1625	0.2265	0.9565
0.1525	25.0102	1690	0.4236	0.9130
0.0002	26.0102	1755	0.7375	0.9130
0.0001	27.0102	1820	1.1523	0.8261
0.0005	28.0102	1885	1.7289	0.7826
0.0237	29.0102	1950	0.7850	0.9130
0.2685	30.0102	2015	1.7432	0.7391
0.0293	31.0102	2080	0.8255	0.9130
0.0085	32.0102	2145	2.1138	0.7391
0.1503	33.0102	2210	0.4744	0.9130
0.0001	34.0102	2275	0.8058	0.8696
0.0004	35.0102	2340	1.8027	0.7391
0.1144	36.0102	2405	3.4276	0.5652
0.0003	37.0102	2470	0.6332	0.8261
0.1925	38.0102	2535	1.6992	0.6957
0.0001	39.0102	2600	1.2077	0.7826
0.0001	40.0102	2665	1.5120	0.8261
0.0	41.0102	2730	1.2723	0.7391
0.063	42.0102	2795	3.7744	0.5652
0.0001	43.0102	2860	3.1736	0.6522
0.1744	44.0102	2925	1.2165	0.8261
0.0001	45.0102	2990	3.7631	0.6087
0.0002	46.0102	3055	0.9466	0.8696
0.0	47.0102	3120	0.9563	0.8261
0.0001	48.0102	3185	1.0371	0.8261
0.1871	49.0102	3250	3.0532	0.6522
0.0036	50.0102	3315	2.5526	0.6522
0.0001	51.0102	3380	1.6451	0.8261
0.2135	52.0102	3445	0.4032	0.9565
0.0	53.0102	3510	0.8310	0.9130
0.0001	54.0102	3575	1.1302	0.8696
0.0001	55.0102	3640	1.7198	0.7826
0.0	56.0102	3705	1.0882	0.8696
0.0	57.0102	3770	1.1533	0.8696
0.0	58.0102	3835	1.4102	0.7826
0.0	59.0102	3900	1.4189	0.7826
0.0	60.0102	3965	1.3660	0.8261
0.0	61.0102	4030	1.2121	0.8696
0.0	62.0102	4095	1.8170	0.7826
0.0	63.0102	4160	1.0369	0.8696
0.0	64.0102	4225	1.3311	0.8696
0.233	65.0102	4290	0.8931	0.9130
0.0	66.0102	4355	0.4937	0.9565
0.0	67.0102	4420	1.5024	0.8261
0.0	68.0102	4485	1.5160	0.8261
0.0	69.0102	4550	1.4395	0.8261
0.0	70.0102	4615	1.2094	0.8261
0.0	71.0102	4680	1.0851	0.8261
0.0	72.0102	4745	1.1003	0.8261
0.0	73.0102	4810	1.0922	0.8261
0.0	74.0102	4875	1.2564	0.8261
0.0	75.0102	4940	1.2901	0.8261
0.0	76.0102	5005	1.2565	0.8261
0.0	77.0102	5070	1.2773	0.8261
0.0	78.0102	5135	1.2317	0.8261
0.0	79.0102	5200	1.2031	0.8261
0.0	80.0102	5265	1.1797	0.8261
0.0	81.0102	5330	1.1727	0.8261
0.0	82.0102	5395	1.1057	0.8261
0.0	83.0102	5460	1.0934	0.8261
0.0	84.0102	5525	1.0780	0.8261
0.0	85.0102	5590	1.0593	0.8261
0.0	86.0102	5655	1.0216	0.8261
0.0	87.0102	5720	1.0498	0.8261
0.0	88.0102	5785	1.0506	0.8261
0.0	89.0102	5850	1.0408	0.8261
0.0	90.0102	5915	1.9110	0.7826
0.0	91.0102	5980	0.7362	0.8696
0.0007	92.0102	6045	1.5211	0.8261
0.0	93.0102	6110	1.2596	0.8696
0.0	94.0102	6175	1.2852	0.8696
0.0	95.0102	6240	1.3083	0.8696
0.0	96.0102	6305	1.3085	0.8696
0.0	97.0102	6370	1.3091	0.8696
0.0	98.0047	6400	1.3091	0.8696

Framework versions

• Transformers 4.46.2
• Pytorch 2.0.1+cu117
• Datasets 3.0.1
• Tokenizers 0.20.0