-
Notifications
You must be signed in to change notification settings - Fork 2
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Can't get it to work. Just getting blurred image right after clicking the icon on the address bar #25
Comments
Some additional info: Not sure if it's hardware related or the recent gpu driver itself since I also tried the cpu version and got the same results - the add-on just blurs the image and doesn't return any translation. I tried waiting for over 5 minutes and also tried refreshing the tab while the icon is still green. |
Hi @himouto thanks for opening this issue. Would it be possible for you to attach also the log shown in the server window? Need to understand what is going on as it is kind of weird that also tesseract would give you the same error. Also i assume you are running using the release file. (I see you can see the advanced options tab in the extension and languages written in extended format so it should be the latest version). Is this your first time trying to run the server or did it ever work for you before with previous versions? |
I will have to investigate and try to reproduce the bug, in my case starting from a fresh install of the server (using the release file v0.5.1) it seems to be working fine. It is kind of weird you don't even see the beginning of the translation starting, but it shouldn't be a a firewall issue since you are able to set the language and models properly One thing you could try is delete the PS: that website does have images that have place-ins replaces by JS when you scroll over them, i still have to implement handling of this in the extension as well |
PS: This would really help, could you give me the result of the following (see screenshot for help, in my case i simulated a server error by not loading any models):
|
Can't see a response in the console. I did a fresh re-install of the add-on and it's still the same - after "GET / HTTP/1.1 200 3776" line in the server window it gets stuck there. I'm using the latest nvidia driver which just updated 2 days ago if that changes anything and Python 3.10.6 for my A1111 stable diffusion setup. Also I can access http://127.0.0.1:4000 just fine so it isn't a local network problem |
There's no >response under Object. I only see >config, >request and >prototype. In the server window it's stuck on "django.server.basehttp - "GET / HTTP/1.1"" - tried waiting it out for over 10 minutes. I have a decent hardware RTX 3070 with the latest driver so it shouldn't take that long. |
So I can tell it' not a problem with
as i am running the same. Also no point in waiting, once there is a red corner it means there was an error (only blurred with no border means it is still doing something) From what you have shown, it looks like the popup.js script is working fine (the one that is doing the setting of the language and models) but the content.js is not (the one running in the activated tab and grabbing and sending images to the server). Weirdly enough you don't even get the URL in the request you have shown (I can simulate the same error by running the translation without the server open, but i see the URL where the image would be sent) I think something is blocking communication between the I will check this evening what could be causing this and get back at you |
If you don't mind could you right click on the trace object and click Unfortunately I am not sure the problem is what i was thinking before. Unfortunately without having the system under my hand or being able to reproduce the bug it is kind of hard to get to the root of it. On a side note, other version of python and installed packages you have on your system should not affect the server regardless as it is made with a frozen installer (PyInstaller) and comes with it's own version of python and dependencies already installed (Might change this in the future as it makes having a plugin manager where users a difficult problem) |
Figured it out. It was uBlock Origin, I disabled it and restarted the server window and it's now working properly Now to figure out why it's doing this... I tried whitelisting 127.0.0.1 and 127.0.0.1:4000 on ublock's Trusted sites but it's still breaking the extension I'm getting these errors on console when ublock is enabled.. not sure if they're related |
Good that you figured it out. The other errors you've shown are not related. One possible long term solution would be for me to modify the extension so all XHR requests are originating either from the For now I will mark this as closed and add a note in the README. Feel free to open a new issue or rewrite here in case something else comes up. Hope you will enjoy the extension |
Thank you for the help. It's a very good extension. |
I'm using run_server-gpu version. The fields are color green but when I click the icon on the address bar nothing happens just blurred image.
I'm getting this error when using kha-white/manga_ocr_base: "Could not find image processor class in the image processor config or the model config. Loading based on pattern matching with the model's feature extractor configuration." so I tested tesseract and the results are the same.
I tested the cpu version and same results - after clicking the add-on icon on the address bar nothing happens it just blurs the
image. Tried refreshing the tab while the icon is green still no translation - just blurred image with red outline. Check screenshot below:
The text was updated successfully, but these errors were encountered: