Open Thunder003 opened 2 years ago
you can first disable the gt encoder and try to overfit one sample to see if your training pipeline works well. BTW, your mask seems not corresponding the input image.
On Sat, Aug 27, 2022 at 8:43 PM Aditya Pandey @.***> wrote:
Hi,
Firstly, thanks for the release.
I've tried on my custom dataset, but the model is not learning and loss taking insane high value. In the end of 40k epochs training got finished due to early stopping & I got :- Train Loss =-7823, Val_Loss= -46939
This is the result:-
[image: DIS] https://user-images.githubusercontent.com/51915348/187056346-33eda7cf-9d6e-4e3e-bded-ac0c09112935.png
In my training dataset, there are around 7k images for vehicles. I trained without using the pre-trained model.
Any Suggestions? Thanks in Advance
— Reply to this email directly, view it on GitHub https://github.com/xuebinqin/DIS/issues/34, or unsubscribe https://github.com/notifications/unsubscribe-auth/ADSGOROEB4FDYWK2CFSZOITV3LN7RANCNFSM572KZZPA . You are receiving this because you are subscribed to this thread.Message ID: @.***>
-- Xuebin Qin PhD Department of Computing Science University of Alberta, Edmonton, AB, Canada Homepage: https://xuebinqin.github.io/
@xuebinqin thanks, I was able to get a proper result.
I'm planning to test on the custom dataset. Can you please tell me which tool you have used to produce the DIS dataset?
GIMP (https://www.gimp.org/) as we described in our paper.
On Sat, Sep 24, 2022 at 10:01 PM Aditya Pandey @.***> wrote:
@xuebinqin https://github.com/xuebinqin thanks, I was able to get a proper result.
I'm planning to test on the custom dataset. Can you please tell me which tool you have used to produce the DIS https://xuebinqin.github.io/dis/index.htmldataset?
— Reply to this email directly, view it on GitHub https://github.com/xuebinqin/DIS/issues/34#issuecomment-1257120797, or unsubscribe https://github.com/notifications/unsubscribe-auth/ADSGORIBRHRQOOG6UAGHAADV77MDZANCNFSM572KZZPA . You are receiving this because you were mentioned.Message ID: @.***>
-- Xuebin Qin PhD Department of Computing Science University of Alberta, Edmonton, AB, Canada Homepage: https://xuebinqin.github.io/
Hi,
Firstly, thanks for the release.
I've tried on my custom dataset, but the model is not learning and loss taking insane high value. In the end of 40k epochs training got finished due to early stopping & I got :- Train Loss =-7823, Val_Loss= -46939
This is the result:-
In my training dataset, there are around 7k images for vehicles. I trained without using the pre-trained model.
Any Suggestions? Thanks in Advance