Hi Like, Very nice video as usual. Can you please make a video on how to fine tune Masked autoencoder on any dataset for denoising downstream task? That really would be very helpful for me at least.
@LukeDitriaАй бұрын
Are you referring to the ViT based Masked autoencoder?
@akramsalim9706Ай бұрын
@@LukeDitria yes exactly
@LukeDitriaАй бұрын
Huggingface does have an implementation of the ViTMAE with pre-trained weights. Though I'll add it to the list of possible videos
@akramsalim9706Ай бұрын
@@LukeDitria thank you very much
@info-md5wjАй бұрын
hey luke great video. Could you please make a video about DeepFill v2? I'm currently working on a magic eraser but am having problems implementing it 🙁
@LukeDitriaАй бұрын
There is quite a lot of code available for DeepFill online. Also, this method is a bit old. Have you had a look at diffusion based infill?
@info-md5wjАй бұрын
@@LukeDitria Yes, but I'm looking for a resource-saving way to carry out inpainting as realistically as possible. Do you have any suggestions as to what would be better than deepfill v2?
@LukeDitriaАй бұрын
You can use DeepFill, though I would recommend starting off with an existing, working implementation.