waifu diffusion tagger V2 traceback

[gcoox]

pls,fix this problem,i try to play it it jump out,traceback cuda failed,is the tagger updating problem??

[konokomi]

yep, it seems there is an error with the tagger since V3 is out. So V2 is an error, need to fix it and add the V3 tagger.

[gcoox]

yep, it seems there is an error with the tagger since V3 is out. So V2 is an error, need to fix it and add the V3 tagger.

ya,any deal like some code to settle this？

[konokomi]

change the keras model

model = keras.layers.TFSMLayer(
args.model_dir,
call_endpoint="serving_default",

or

tag_images_by_wd14_tagger.zip

change the file on google colab

[gcoox]

change the keras model

model = keras.layers.TFSMLayer( args.model_dir, call_endpoint="serving_default",

or

tag_images_by_wd14_tagger.zip

change the file on google colab

is work!!!but next time pls tell where's destination file have to drop/change ,i have been stumped in this moment a while ,but anywhere thx buddy :D YOU ARE GENIUS [[for some people who dont know the file drop in colab : go to kohya trailer>finetuner>drop here]] P.S. but if is change keras model i dont know how TO DOXP

[PlumButa]

It's not work. The zip file is empty. Did I download it the wrong way?🤔

I tried to modify Linaqruf's "tag_images_by_wd14_tagger.py" and upload it to colab, but it was not successful (I am not good at coding😂).

If anyone could explain in more detail how to modify the code to make it work, I'd be very grateful.

[DEX-1101]

actually just update the xformers and it will fix the tagger issue

[klingki]

It's not work. The zip file is empty. Did I download it the wrong way?🤔

I tried to modify Linaqruf's "tag_images_by_wd14_tagger.py" and upload it to colab, but it was not successful (I am not good at coding😂).

If anyone could explain in more detail how to modify the code to make it work, I'd be very grateful.

I just figured this out. Unzip the file with 7zip, not your built-in unzipper. There should be two .py files inside. Inside the colab, click on Files (icon shaped like a folder, all the way on the left). Then drop the 2 files in the location gcoox said above (kohya-trainer>finetune).

After trying a million different things, this finally worked for me. Hopefully the colab just gets updated soon so we don't have to mess with this stuff (I know nothing at all about python/coding 😭)