videomae-base-finetuned-dogBehavior

<!-- 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-base-finetuned-dogBehavior

This model is based on the VideoMae base model and was finetuned on videos of sniffer dogs detecting Parkinson's disease.

The dogs are trained to lie down when they detect Parkinson's disease in a sample. The training is done with a mix of negative and positive samples.

It achieves the following results on the evaluation set:

• Loss: 0.0611
• Accuracy: 0.9910

Model description

The dataset consists of video clips of 2 behavior types : lying down and exploratory sniffing the sample. The goal of the model is to detect other behavior.

Intended uses & limitations

Weights are uploaded but video clips or images are protected and cannot be shared.

Training and evaluation data

The model was trained with 25 epochs. If the dataset changes, the model is retrained for 10 epochs. The accuracy to detect the behavior of lying and sniffing is 99%. Detections that are under this confidence are considered as "other" behavior that needs further analysis in respect ot its relation to the type of sample (negative or positive).

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 5e-05
• train_batch_size: 8
• eval_batch_size: 8
• seed: 42
• optimizer: Use adamw_torch_fused 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: 370

Training results

Framework versions

• Transformers 4.57.3
• Pytorch 2.9.1+cu130
• Datasets 4.4.2
• Tokenizers 0.22.2

Sources

basemodel: MCG-NJU/videomae-base

misc{https://doi.org/10.48550/arxiv.2203.12602, doi = {10.48550/ARXIV.2203.12602}, url = {https://arxiv.org/abs/2203.12602}, author = {Tong, Zhan and Song, Yibing and Wang, Jue and Wang, Limin}, keywords = {Computer Vision and Pattern Recognition (cs.CV), FOS: Computer and information sciences, FOS: Computer and information sciences}, title = {VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training}, publisher = {arXiv}, year = {2022}, copyright = {Creative Commons Attribution 4.0 International} }

Many thanks to

Chang-Qing Gao and Ni Wang and Meimei Wang et al., "Sensitivity of Sniffer Dogs for a Diagnosis of Parkinson's Disease: A Diagnostic Accuracy Study" Aug. 2022. Lab: Chang-Qing Gao's Lab [Online].Available: {https://www.researchgate.net/publication/362936321_Sensitivity_of_Sniffer_Dogs_for_a_Diagnosis_of_Parkinson's_Disease_A_Diagnostic_Accuracy_Study}

Howest University for Applied Sciences, Creative Technologies and Artificial Intelligence