mayjs
mayjs
Unfortunately I don't have any experience with LanguageTools Java API. I found this rule [in the list of error detection rules](https://community.languagetool.org/rule/show/PASSIVE_VOICE?lang=en&subId=5). Doesn't the API provide a way to enable rules...
I think the [enableRule](https://languagetool.org/development/api/org/languagetool/JLanguageTool.html#enableRule-java.lang.String-) method would allow you to enable the rule by ID. Ideally this would be exposed to a command line argument.
Thanks for your answer! I finally had some time to take a look at this again. You are right, the StackOverflow answer contains exactly the right snippet. For reference, if...
Hey, I went ahead and extended your branch with an implementation of `mode=reflect` in the shader :) You can take a look at it here: https://github.com/webonnx/wonnx/commit/866a5c4e3f314a0db140fd209698b435713f3e51 I'm not sure if...
Okay, I'll try to explain the most important parts, let me know if you need some more details. The most essential thing you need to understand about WGSL is the...
I think you just forgot to adjust the "X" input size for your new input tensor dimensions. It should probably be `&[3, 3]` right? Regarding your output values: I think...
> I finally got around to fixing it, lol. @mayjs you implied that there might be another bug? Hey, sorry for the late answer, I'm glad that you fixed it...
Okay, I think we have the most important features of `ConvTranspose` supported now. We are missing support for 1D and 3D, `output_padding`, `auto_pad`, `dilations`, and `group`. I'm also not sure...
Also, as I mentioned before, this was the missing step for me to run UNet. I have successfully used this to run a pretrained [nind-denoise](https://github.com/trougnouf/nind-denoise) model. We could at a...
@pixelspark Hm, looking through the log files (https://github.com/webonnx/wonnx/actions/runs/5843079870/job/15844866120?pr=182), ConvTranspose is not listed in the loaded / supported operations. Not sure where the operations are registered, but it's looking like the...