videoMAE_base_wlasl_100_50ep_coR_p10

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: 2.5394
• Accuracy: 0.4911

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: 2
• eval_batch_size: 2
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 8
• 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: 9000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
18.6484	0.02	180	4.6434	0.0089
18.5863	1.0200	360	4.6257	0.0178
18.5141	2.0199	540	4.6094	0.0237
18.3951	3.0200	721	4.6105	0.0207
18.4598	4.02	901	4.6059	0.0148
18.3257	5.0200	1081	4.5610	0.0266
17.9377	6.0199	1261	4.4575	0.0325
17.0809	7.0200	1442	4.3142	0.0355
16.4349	8.02	1622	4.2145	0.0296
15.8097	9.0200	1802	3.9967	0.0592
14.6937	10.0199	1982	3.9020	0.0917
13.3281	11.0200	2163	3.6937	0.1036
12.0028	12.02	2343	3.4095	0.1834
10.4545	13.0200	2523	3.1194	0.2574
8.9317	14.0199	2703	3.0132	0.2604
7.3575	15.0200	2884	2.8562	0.2959
6.0437	16.02	3064	2.6942	0.3373
4.8818	17.0200	3244	2.5883	0.3550
3.8135	18.0199	3424	2.5386	0.3609
2.7936	19.0200	3605	2.4362	0.4053
2.1059	20.02	3785	2.4949	0.4260
1.5309	21.0200	3965	2.3770	0.3935
1.0911	22.0199	4145	2.3148	0.4467
0.8395	23.0200	4326	2.3940	0.4231
0.6525	24.02	4506	2.2907	0.4675
0.4313	25.0200	4686	2.2253	0.4615
0.3262	26.0199	4866	2.2776	0.4852
0.2097	27.0200	5047	2.4151	0.4556
0.1815	28.02	5227	2.3768	0.4793
0.1124	29.0200	5407	2.4308	0.4852
0.1165	30.0199	5587	2.4429	0.5
0.0915	31.0200	5768	2.4644	0.4822
0.0822	32.02	5948	2.5156	0.4911
0.0555	33.0200	6128	2.6018	0.4941
0.0693	34.0199	6308	2.5839	0.4822
0.0609	35.0200	6489	2.6524	0.4882
0.0719	36.02	6669	2.6280	0.4793
0.067	37.0200	6849	2.7113	0.4675
0.0592	38.0199	7029	2.5907	0.4822
0.0621	39.0200	7210	2.6348	0.4793
0.0791	40.02	7390	2.5394	0.4911

Framework versions

• Transformers 4.46.1
• Pytorch 2.5.1+cu124
• Datasets 3.1.0
• Tokenizers 0.20.1