Aller au contenu
ModelPrésentéExterne
Télécharger

VideoMAE Kinetics Wlasl 100 Signer 200ep 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: 1.4810 Accuracy: 0.7071

Sourcehuggingface.co/Shawon16/videoMAE_kinetics_wlasl_100__signer_200ep_coR/tree/bb393d659a16037de772b3d9d843546a01fb9106
TypeOtherModèle de baseMCG-NJU/videomae-base-finetuned-kineticsVersionRevision bb393d6CClasse

Source du modèle

Type
Other
Taille du fichier
329 MB
Modèle de base
MCG-NJU/videomae-base-finetuned-kinetics
Version
vRevision bb393d6
Mis à jour
24 oct. 2026
Téléchargements
11
Source
Hugging Face

Description de la source

videoMAE_kinetics_wlasl_100__signer_200ep_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: 1.4810
  • Accuracy: 0.7071

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 LossEpochStepValidation LossAccuracy
18.5250.0051804.61330.0148
18.4261.00503604.58440.0355
18.08922.00505404.48540.0473
16.92673.00507214.20440.1095
15.40064.0059013.82100.2367
13.38735.005010813.46360.3402
11.39716.005012613.06290.4349
9.32437.005014422.72340.5148
7.4888.00516222.37140.5799
5.74299.005018022.03920.6450
4.270310.005019821.84130.6568
3.035111.005021631.60960.6627
2.119112.00523431.45380.6982
1.399713.005025231.36700.6627
0.955214.005027031.33830.6953
0.641815.005028841.21760.7189
0.47616.00530641.16030.7071
0.341317.005032441.17710.7041
0.252918.005034241.04120.7278
0.225519.005036051.24960.6746
0.167520.00537851.19290.7101
0.109321.005039651.19920.7041
0.113822.005041451.23770.7189
0.143223.005043261.22300.7071
0.086824.00545061.19350.7308
0.080325.005046861.22110.7160
0.131226.005048661.29530.6953
0.135627.005050471.42100.6746
0.133828.00552271.20000.7219
0.113229.005054071.36020.6953
0.126630.005055871.36820.6923
0.101931.005057681.20610.7189
0.11332.00559481.32690.7041
0.184633.005061281.36740.6893
0.133734.005063081.41510.7101
0.104535.005064891.29970.7249
0.141936.00566691.48810.6953
0.101537.005068491.45750.7012
0.098438.005070291.44810.6864
0.147839.005072101.89200.6598
0.221540.00573901.63550.6982
0.182741.005075701.63370.6805
0.075842.005077501.58140.6893
0.10943.005079311.52560.7012
0.053944.00581111.48100.7071

Framework versions

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

Sources

1 source
Hugging FacePrincipalVersion bb393d659a16037de772b3d9d843546a01fb9106
huggingface.co/Shawon16/videoMAE_kinetics_wlasl_100__signer_200ep_coR/tree/bb393d659a16037de772b3d9d843546a01fb9106

Vérifié 24 oct.

Artefacts du modèle

1 artefact

model.safetensors

safetensors · 329 MB · SHA-256 3e0004532c7a…d140 · Hugging Face

Download file

Extraits de sources

3 extraits
Readme6 aoûthuggingface.co/Shawon16/videoMAE_kinetics_wlasl_100__signer_200ep_coR

--- library_name: transformers license: cc-by-nc-4.0 base_model: MCG-NJU/videomae-base-finetuned-kinetics tags: - generated_from_trainer metrics: - accuracy model-index: - name: videoMAE_kinetics_wlasl_100__signer_200ep_coR results: [] --- <!-- This model card has been generated automatically according to the information the Trainer had access to. You should probably proofread and complete it, then remove this comment. --> # videoMAE_kinetics_wlasl_100__signer_200ep_coR This model is a fine-tuned version of [MCG-NJU/videomae-base-finetuned-kinetics](https://huggingface.co/MCG-NJU/videomae-base-finetuned-kinetics) on an unknown dataset. It achieves the following results on the evaluation set: - Loss: 1.4810 - Accuracy: 0.7071 ## 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 | Accuracy | |:-------------:|:-------:|:----:|:---------------:|:--------:| | 18.525 | 0.005 | 180 | 4.6133 | 0.0148 | | 18.426 | 1.0050 | 360 | 4.5844 | 0.0355 | | 18.0892 | 2.0050 | 540 | 4.4854 | 0.0473 | | 16.9267 | 3.0050 | 721 | 4.2044 | 0.1095 | | 15.4006 | 4.005 | 901 | 3.8210 | 0.2367 | | 13.3873 | 5.0050 | 1081 | 3.4636 | 0.3402 | | 11.3971 | 6.0050 | 1261 | 3.0629 | 0.4349 | | 9.3243 | 7.0050 | 1442 | 2.7234 | 0.5148 | | 7.488 | 8.005 | 1622 | 2.3714 | 0.5799 | | 5.7429 | 9.0050 | 1802 | 2.0392 | 0.6450 | | 4.2703 | 10.0050 | 1982 | 1.8413 | 0.6568 | | 3.0351 | 11.0050 | 2163 | 1.6096 | 0.6627 | | 2.1191 | 12.005 | 2343 | 1.4538 | 0.6982 | | 1.3997 | 13.0050 | 2523 | 1.3670 | 0.6627 | | 0.9552 | 14.0050 | 2703 | 1.3383 | 0.6953 | | 0.6418 | 15.0050 | 2884 | 1.2176 | 0.7189 | | 0.476 | 16.005 | 3064 | 1.1603 | 0.7071 | | 0.3413 | 17.0050 | 3244 | 1.1771 | 0.7041 | | 0.2529 | 18.0050 | 3424 | 1.0412 | 0.7...

Note24 oct.huggingface.co/Shawon16/videoMAE_kinetics_wlasl_100__signer_200ep_coR

Source context: 11 downloads · 0 likes · Pipeline video-classification · Library transformers · Repo Shawon16/videoMAE_kinetics_wlasl_100__signer_200ep_coR

Note6 aoûthuggingface.co/Shawon16/videoMAE_kinetics_wlasl_100__signer_200ep_coR

Source context: 2 downloads · 0 likes · Pipeline video-classification · Library transformers · Repo Shawon16/videoMAE_kinetics_wlasl_100__signer_200ep_coR