Modelpublic

Shawon16/videoMAE_kinetics_wlasl_100__signer_20ep_coR

sourceHugging Facecc-by-nc-4.0updated 10mo agoView on Hugging Face

0likes66downloads

Model Card

videoMAEkineticswlasl100signer20ep_coR

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

—Loss: 0.8901
—Accuracy: 0.7811

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
—trainbatchsize: 2
—evalbatchsize: 2
—seed: 42
—gradientaccumulationsteps: 4
—totaltrainbatch_size: 8
—optimizer: Use adamwtorch with betas=(0.9,0.999) and epsilon=1e-08 and optimizerargs=No additional optimizer arguments
—lrschedulertype: linear
—lrschedulerwarmup_ratio: 0.1
—training_steps: 3600
—mixedprecisiontraining: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss	Accuracy
18.4525	0.05	180	4.5296	0.0325
16.671	1.0499	360	3.6968	0.1716
12.2061	2.0499	540	2.8214	0.4290
8.0718	3.0501	721	2.1703	0.5562
5.3342	4.05	901	1.8351	0.6213
3.3842	5.0499	1081	1.5110	0.6923
2.0996	6.0499	1261	1.3504	0.6864
1.253	7.0501	1442	1.1704	0.7249
0.8034	8.05	1622	1.1163	0.7219
0.4457	9.0499	1802	1.0392	0.7367
0.2809	10.0499	1982	1.1198	0.7012
0.1655	11.0501	2163	0.9566	0.7456
0.1146	12.05	2343	0.9404	0.7633
0.0727	13.0499	2523	0.9171	0.7663
0.0646	14.0499	2703	0.8766	0.7722
0.0575	15.0501	2884	0.9062	0.7633
0.0475	16.05	3064	0.9065	0.7663
0.051	17.0499	3244	0.9042	0.7663
0.0406	18.0499	3424	0.8983	0.7751
0.0467	19.0487	3600	0.8901	0.7811

Framework versions

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