In part 2 of this series, we start using the "transformers-cli add-new-model-like" command to quickly prototype a skeleton for the new model. We also start implementing the first layer of the model, which for DINOv2 is the patch embedding layer. UA-cam is still processing the video, but will crispy sharp in a few hours :)
In part 2 of this series, we start using the "transformers-cli add-new-model-like" command to quickly prototype a skeleton for the new model. We also start implementing the first layer of the model, which for DINOv2 is the patch embedding layer.
UA-cam is still processing the video, but will crispy sharp in a few hours :)
Thanks a lot
This is gold