search

plemeri / InSPyReNet

Official PyTorch implementation of Revisiting Image Pyramid Structure for High Resolution Salient Object Detection (ACCV 2022)
MIT License
320 stars 61 forks source link

Fine tuning #44

Open rishabh063 opened 5 months ago

rishabh063 commented 5 months ago

Hey I have been fine tuning using pre-trained checkpoint . however Having issues for it to converge , I have 20k Mask and image set for background removal . I have run 30 Epochs and the results are worse than the pretrained one

Any advice on how many epoch or learning rate ?

plemeri commented 5 months ago

Finetuning with additional data only may cause catastrophic forgetting. Try adding your data to the composite dataset.

LukaGiorgadze commented 5 months ago

Finetuning with additional data only may cause catastrophic forgetting. Try adding your data to the composite dataset.

I'm just curious - what's the reason of that?

rishabh063 commented 5 months ago

So you mean i Need to train on the pretraining dataset and my own dataset ? Can this be an issue with my dataset?

rishabh063 commented 4 months ago

@plemeri I saw that in your transparent background libraries nightly model scores much higher than HR model in model zoo when tested on DUTS-TE . Can you share the reason and training technique ? is it just trained longer ?

plemeri commented 4 months ago

Nightly model was trained with the updated training strategy which was my doctorate dissertation. Currently I have no time to release updated contents.

rishabh063 commented 4 months ago

can you share that dissertation document ? i can update the code

rishabh063 commented 4 months ago

@plemeri if you can share it

plemeri commented 4 months ago

My dissertation is not publicly opened yet because of the ongoing patent registration. Sorry for that.

LukaGiorgadze commented 4 months ago

@plemeri congrats on your doctorate dissertation! 👍 I would be more than happy to see the updates on this repository and not kill as everyone does after some time. We are waiting updated training strategy at some point.