Hi I wanted to try out the selfsupervision based Loca project but only image net is supported. I would like to use the model to be pretrained on my own small dataset (32k images) and see some results. Or would you suggest finetuning it for segmentation in colab.
There does not appear to be any documentation for making a custom dataset pipeline for intergration into already existing project.
can you let me know about it ?
Lastly how to convert these weights to maybe Pytorch to host it on Huggingface will be a great addition.
Hi I wanted to try out the selfsupervision based Loca project but only image net is supported. I would like to use the model to be pretrained on my own small dataset (32k images) and see some results. Or would you suggest finetuning it for segmentation in colab.
There does not appear to be any documentation for making a custom dataset pipeline for intergration into already existing project.
can you let me know about it ?
Lastly how to convert these weights to maybe Pytorch to host it on Huggingface will be a great addition.