DenseNet and ReseNet PyTorch blink models

Tobias-Fischer / rt_gene

RT-GENE: Real-Time Eye Gaze and Blink Estimation in Natural Environments

http://www.imperial.ac.uk/personal-robotics

Other

374 stars 71 forks source link

DenseNet and ReseNet PyTorch blink models #105

Closed SohilZidan closed 3 years ago

SohilZidan commented 3 years ago

Hello,

where I can find the pretrained models on DenseNet and ResNet backbones? I am trying to use estimate_blink_standalone.py

Best

Tobias-Fischer commented 3 years ago

Hiya,

Currently we do not have pretrained models (unless you have some @ahmed-alhindawi?). You can easily use the train_model.py in the rt_bene_model_training/pytorch folder and change the model_base to train a model yourself though: https://github.com/Tobias-Fischer/rt_gene/blob/6eb7c32ff123a3f7b13bc9a492080a72158b14b0/rt_bene_model_training/pytorch/train_model.py#L84-L86

Closing here - let us know if you have more questions.

ahmed-alhindawi commented 3 years ago

Sorry, there aren't any publically available Pytorch models for DenseNet or ResNet at the moment. But Tobias is right - it'll be easy to train if you'd like to.

SohilZidan commented 3 years ago

can you provide versioned requirements? @Tobias-Fischer @ahmed-alhindawi