This would be a really nice feature in darknet without having to go via pytorch or keras. In doing so you sometimes loose fine elements of the model if other repositories don鈥檛 replicate the model exactly in their framework of choice.
All 2 comments
This would be a really nice feature in darknet without having to go via pytorch or keras. In doing so you sometimes loose fine elements of the model if other repositories don鈥檛 replicate the model exactly in their framework of choice.
Most helpful comment
This would be a really nice feature in darknet without having to go via pytorch or keras. In doing so you sometimes loose fine elements of the model if other repositories don鈥檛 replicate the model exactly in their framework of choice.