videomae-base-ipm_all_videos_gb

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: 1.2748
• Accuracy: 0.6870

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: 4
• eval_batch_size: 4
• 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: 4800

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
2.5051	0.01	60	2.5234	0.0870
2.4957	1.01	120	2.5401	0.1217
2.5475	2.01	180	2.5675	0.0870
2.4659	3.01	240	2.5836	0.0957
2.2644	4.01	300	2.5035	0.0696
2.3548	5.01	360	2.2569	0.1217
2.0341	6.01	420	2.3958	0.1565
2.2919	7.01	480	2.6096	0.0696
2.0857	8.01	540	2.3223	0.1217
1.7473	9.01	600	2.5414	0.1652
1.885	10.01	660	1.7822	0.3043
1.9496	11.01	720	1.8052	0.3130
1.2315	12.01	780	2.1955	0.2435
1.3549	13.01	840	2.1262	0.3130
1.5121	14.01	900	2.0316	0.2783
1.4504	15.01	960	1.7596	0.2957
1.2991	16.01	1020	1.6413	0.3652
1.2299	17.01	1080	1.5417	0.4087
1.2965	18.01	1140	1.7243	0.3739
1.2431	19.01	1200	1.7556	0.3478
1.3807	20.01	1260	1.4580	0.4435
1.3961	21.01	1320	1.6514	0.4
1.0119	22.01	1380	1.5449	0.3391
1.3799	23.01	1440	1.5126	0.3304
1.6871	24.01	1500	2.0675	0.2783
1.2707	25.01	1560	1.7128	0.3739
1.1495	26.01	1620	1.6387	0.3217
1.6151	27.01	1680	1.6192	0.3913
1.0587	28.01	1740	1.6008	0.4522
1.2169	29.01	1800	1.6739	0.4348
1.1116	30.01	1860	1.7693	0.3913
1.0939	31.01	1920	1.6540	0.3913
0.9307	32.01	1980	1.5583	0.4957
0.9539	33.01	2040	1.8836	0.4174
0.9804	34.01	2100	1.5656	0.4522
1.334	35.01	2160	1.5375	0.4609
1.0897	36.01	2220	1.4327	0.4087
0.864	37.01	2280	1.6372	0.3913
0.9678	38.01	2340	1.4537	0.4609
1.3184	39.01	2400	1.3085	0.4783
1.1462	40.01	2460	1.4954	0.4696
0.7875	41.01	2520	1.4692	0.4870
0.9552	42.01	2580	1.3797	0.4174
0.8053	43.01	2640	1.3216	0.5043
0.9231	44.01	2700	1.2134	0.5739
0.734	45.01	2760	1.1676	0.5304
0.5427	46.01	2820	1.2179	0.4783
0.7171	47.01	2880	1.2749	0.5304
0.6977	48.01	2940	1.3707	0.5304
0.6911	49.01	3000	1.2520	0.5478
0.6166	50.01	3060	1.3687	0.5304
0.4025	51.01	3120	1.4041	0.5652
0.6147	52.01	3180	1.3030	0.6435
0.5787	53.01	3240	1.4109	0.5913
0.7157	54.01	3300	1.3183	0.6
0.3391	55.01	3360	1.4333	0.5913
0.7482	56.01	3420	1.4549	0.5826
0.5182	57.01	3480	1.4193	0.5652
0.7383	58.01	3540	1.4043	0.5565
0.8862	59.01	3600	1.4041	0.6
0.3481	60.01	3660	1.3164	0.6435
0.763	61.01	3720	1.2947	0.5913
0.7397	62.01	3780	1.2785	0.6696
0.514	63.01	3840	1.3180	0.6522
0.6582	64.01	3900	1.3520	0.6696
0.3929	65.01	3960	1.3391	0.6609
0.7623	66.01	4020	1.4349	0.6348
0.6235	67.01	4080	1.2897	0.6522
0.449	68.01	4140	1.3150	0.6696
0.639	69.01	4200	1.4241	0.6087
0.473	70.01	4260	1.2578	0.6609
0.5478	71.01	4320	1.2770	0.6522
0.4732	72.01	4380	1.2893	0.6783
0.5489	73.01	4440	1.2195	0.7043
0.3907	74.01	4500	1.2523	0.6957
0.2572	75.01	4560	1.2149	0.7043
0.5022	76.01	4620	1.2934	0.6696
0.2958	77.01	4680	1.2726	0.6783
0.7009	78.01	4740	1.2779	0.6957
0.49	79.01	4800	1.2748	0.6870

Framework versions

• Transformers 4.29.1
• Pytorch 2.0.1+cu117
• Datasets 2.12.0
• Tokenizers 0.13.3