VideoMAE_Base_WLASL_100_200_epochs_p20_SR_8

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: 2.8301
• Top 1 Accuracy: 0.5296
• Top 5 Accuracy: 0.7988
• Top 10 Accuracy: 0.8698
• Accuracy: 0.5296
• Precision: 0.5768
• Recall: 0.5296
• F1: 0.5133

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: 36000
• mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Top 1 Accuracy	Top 5 Accuracy	Top 10 Accuracy	Accuracy	Precision	Recall	F1
18.5755	0.005	180	4.6413	0.0089	0.0444	0.1006	0.0089	0.0005	0.0089	0.0010
18.4843	1.0050	360	4.6308	0.0089	0.0621	0.1006	0.0089	0.0006	0.0089	0.0012
18.4592	2.0050	541	4.6209	0.0207	0.0710	0.1065	0.0207	0.0017	0.0207	0.0027
18.4604	3.005	721	4.6171	0.0178	0.0680	0.1213	0.0178	0.0005	0.0178	0.0010
18.4543	4.0050	901	4.6106	0.0237	0.0858	0.1450	0.0237	0.0008	0.0237	0.0016
18.3384	5.0050	1081	4.6190	0.0178	0.0680	0.1095	0.0178	0.0003	0.0178	0.0007
18.1947	6.0050	1262	4.6096	0.0296	0.0917	0.1361	0.0296	0.0038	0.0296	0.0057
18.1104	7.005	1442	4.5992	0.0355	0.0947	0.1450	0.0355	0.0060	0.0355	0.0082
18.0825	8.0050	1622	4.5940	0.0178	0.0976	0.1568	0.0178	0.0059	0.0178	0.0080
17.9081	9.0050	1802	4.5627	0.0325	0.1065	0.1686	0.0325	0.0076	0.0325	0.0103
17.4986	10.0050	1983	4.4287	0.0355	0.1213	0.2012	0.0355	0.0087	0.0355	0.0126
16.3869	11.005	2163	4.1366	0.0651	0.2337	0.3757	0.0651	0.0228	0.0651	0.0289
15.0278	12.0050	2343	3.7821	0.1065	0.3580	0.5178	0.1065	0.0601	0.1065	0.0572
13.4974	13.0050	2523	3.4419	0.2012	0.5030	0.6686	0.2012	0.1761	0.2012	0.1544
11.5774	14.0050	2704	3.2605	0.2189	0.5355	0.6982	0.2189	0.2069	0.2189	0.1725
10.263	15.005	2884	2.8508	0.3195	0.6627	0.8077	0.3195	0.3730	0.3195	0.2952
8.1589	16.0050	3064	2.5945	0.3905	0.7308	0.8462	0.3905	0.4168	0.3905	0.3592
6.8221	17.0050	3244	2.4311	0.3994	0.7337	0.8728	0.3994	0.4000	0.3994	0.3640
5.4923	18.0050	3425	2.2139	0.4615	0.7929	0.8669	0.4645	0.4886	0.4645	0.4350
4.1619	19.005	3605	2.1384	0.4734	0.7840	0.8817	0.4734	0.5276	0.4734	0.4503
3.3413	20.0050	3785	1.9583	0.5118	0.8107	0.9053	0.5118	0.5485	0.5118	0.4908
2.5832	21.0050	3965	1.8604	0.5	0.8284	0.9201	0.5	0.5000	0.5	0.4713
1.9003	22.0050	4146	1.9390	0.5296	0.8195	0.8905	0.5296	0.5860	0.5296	0.5140
1.4226	23.005	4326	1.9288	0.5266	0.8077	0.9083	0.5266	0.6341	0.5266	0.5303
1.1341	24.0050	4506	1.8854	0.5266	0.7899	0.8964	0.5266	0.5719	0.5266	0.5079
0.9315	25.0050	4686	1.7328	0.5769	0.8462	0.8994	0.5769	0.6261	0.5769	0.5565
0.7536	26.0050	4867	1.8349	0.5385	0.8195	0.8964	0.5385	0.6288	0.5385	0.5296
0.4518	27.005	5047	1.7999	0.5533	0.8550	0.9172	0.5533	0.6276	0.5533	0.5437
0.3322	28.0050	5227	1.6931	0.6006	0.8491	0.9260	0.6006	0.6578	0.6006	0.5896
0.403	29.0050	5407	1.8000	0.5740	0.8462	0.9142	0.5740	0.6240	0.5740	0.5584
0.1837	30.0050	5588	1.8765	0.5769	0.8373	0.8935	0.5769	0.6391	0.5769	0.5664
0.1579	31.005	5768	2.0752	0.5473	0.8432	0.8994	0.5473	0.6337	0.5473	0.5426
0.2079	32.0050	5948	1.9234	0.5947	0.8136	0.8964	0.5947	0.6524	0.5947	0.5791
0.2738	33.0050	6128	1.8529	0.6036	0.8491	0.9024	0.6065	0.6622	0.6065	0.5892
0.2621	34.0050	6309	1.9906	0.5740	0.8284	0.9172	0.5740	0.6191	0.5740	0.5640
0.2024	35.005	6489	1.8942	0.5976	0.8639	0.9260	0.5976	0.6615	0.5976	0.5886
0.0983	36.0050	6669	2.0340	0.5858	0.8254	0.8846	0.5858	0.6500	0.5858	0.5729
0.0592	37.0050	6849	1.8493	0.6095	0.8609	0.9231	0.6095	0.6775	0.6095	0.5986
0.0922	38.0050	7030	1.9036	0.6302	0.8669	0.9260	0.6302	0.6825	0.6302	0.6125
0.1547	39.005	7210	1.9897	0.6036	0.8432	0.9053	0.6036	0.6726	0.6036	0.5948
0.1162	40.0050	7390	2.3056	0.5828	0.8284	0.8876	0.5828	0.6518	0.5828	0.5675
0.0514	41.0050	7570	2.3211	0.5888	0.7988	0.8817	0.5888	0.6510	0.5888	0.5767
0.1138	42.0050	7751	2.3149	0.5740	0.8491	0.9053	0.5740	0.6285	0.5740	0.5628
0.1197	43.005	7931	2.1156	0.6124	0.8669	0.9172	0.6124	0.6907	0.6124	0.6073
0.0673	44.0050	8111	2.2835	0.5828	0.8580	0.9083	0.5828	0.6274	0.5828	0.5641
0.1501	45.0050	8291	2.2719	0.5917	0.8521	0.8905	0.5917	0.6419	0.5917	0.5757
0.2022	46.0050	8472	2.3422	0.5562	0.8491	0.9053	0.5562	0.6034	0.5562	0.5402
0.2185	47.005	8652	2.6431	0.5237	0.8284	0.8817	0.5237	0.5808	0.5237	0.5125
0.2385	48.0050	8832	2.3147	0.5799	0.8521	0.9053	0.5799	0.6324	0.5799	0.5623
0.1769	49.0050	9012	2.3451	0.5769	0.8373	0.8876	0.5769	0.6246	0.5769	0.5622
0.1927	50.0050	9193	2.7140	0.5562	0.8018	0.8728	0.5562	0.6025	0.5562	0.5347
0.2048	51.005	9373	2.3876	0.5917	0.8225	0.8935	0.5917	0.6367	0.5917	0.5748
0.1608	52.0050	9553	2.6983	0.5266	0.8077	0.8580	0.5266	0.5645	0.5266	0.5013
0.1256	53.0050	9733	2.7464	0.5385	0.8018	0.8905	0.5385	0.5773	0.5385	0.5257
0.1327	54.0050	9914	2.5133	0.5651	0.7988	0.8846	0.5651	0.5812	0.5651	0.5342
0.0503	55.005	10094	2.5687	0.5769	0.8314	0.8964	0.5769	0.6336	0.5769	0.5635
0.0841	56.0050	10274	2.7311	0.5503	0.8284	0.8905	0.5503	0.6025	0.5503	0.5329
0.0888	57.0050	10454	2.6771	0.5592	0.8195	0.8935	0.5592	0.6142	0.5592	0.5473
0.0629	58.0050	10635	2.8301	0.5296	0.7988	0.8698	0.5296	0.5768	0.5296	0.5133

Framework versions

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