@orange said in 8 Times more CPU consumption on Aida-X Neural Models:

@DabDab I haven't trained any model yet, but loading models is done like this:

Open the Neural Sine synth example in the Snippet Browser. Delete the synth. Then replace the sine example with one of the models above (the models are in json format, just copy/paste). Then open the neural node in FX as Scriptnode, select the model then you're good to go.

When you replace the neural network, you might need to restart HISE after saving, sometimes it doesn't update.

Wow... i will give it a try. Later I will need @griffinboy help.

@orange Here's an example:

HiseSnippet 1549.3ocuV02SaaDF2FvsKYcSsR6emz8WUtqHZRgBkQaWgPxZzJgHLEUooIzg8Y7Ibtyx2Yfro9+6i09.rOL6av1yycNINjzWVzVopfumW+cOudrq0LQDK5TkfmkwzNNtK2OWFxTJYtia8iGlwbbuiWvPgNoUBkKb5tui6W6c.UoY4DKo8FlQUJVjiq6x+HRvs1JNle9qeXOZJUDxlPxw4DIOj8Z9.tdB09u7m3oocnQri4CpH8FuranTzRlJK.7rrWCmLZ3EzyY8nnXK447JpJww867Z7jMYrsVONJht95Qat4la+zs2X6lz3FaPOq4lauwi2XqG+zshcbuU6HtVlGnoZlxwck8jQCCRjWIrN3DtheVJCOzzI.7rkbGYZDdEQpNsR3oQiCTJHrsRkv1x1v123c.OhOl9jv2cMLHSznZ.zcoog2xSAulUgWiJvaNPxsBjVwBo64EDlyyzS3f34K85JfrYLExSUghUVG2+vqkDDPnWa.8BVmb3vXE72rQiUIvudvNwEhPMWJHRQOolcnv+A0+s50p+t5jaxJNdt7P2jKSSY4ykMVZj+gTzWTL3LV9pjKooErwBBW+oiodu+XZ0Tdn8VWQPonqfqOLiIdeEBNkgJ3q2zceplhIhRZfbYrbMGgf69rKg1.aZol29L0EZYFzHLSNCpVjQEoT8zkPXiVICHFLUdCSNBEWOrZi3+Y0UepP7dd845vj4iwklCFgH0+GXrra7q7ZGGyB0S.3Jdcd6muVu5ksdfOQmemxijNus4ba59cuIE3Y4rLZN6XY+T5PeEcPVJ6H.iqRNKUFdQ.+WYy1UjYgvdnD9gITgfkpVjlma8IGUZ9gSJGIKzbw4GP047qgtvdECBf45grVknCn4tD1CYO2.Oi0EAvRJyg+F9ojYS7raIyliXVo8qGSekL+BS5n7aG2aai8JSP9z3qA3yxUXwo6s8ZrF7OmcSSkW0RNHiWV1BIECs9xzgYIRAODIYkXDz2cfr.7ZI9gcRGS4oXcdPgBlVDcnH.D1rHDqy5IifutUGZHDXG1mpSv9EbdCTbxxWKbbQ5TXsZchcA3TrQqNZUoMvAfb.FMvEkfJ.ea3v5d6rekfc8oBjxTv4t.bxYqMg6rNupteDeuRkYe11C7Dz7uDBhZdsRAa2Mx4Dr1yL6Pp5wNuCOE1y.0LFyAU8WWNR06XJTygEUAo7HVde3UA3ocKh3RPKCKq13W6yUYPqydEvH.67AZNbEJ4NYD0wz7yY5YD3CB9630UcBxMjlNB+PUvdR4ECnl5uEaSvmitkAzvb4og1A.3k6KLTfZEgICWCdnGblzz4R6Eaxr5AvSXNMLbZSMihOdQUb8EUwMVTEexhp3lKphasnJ9zOth3KL2sPKGXmgAE48aaJBccaKvdGSerCCF4GAuq0t0AqUGSAZsKajbci43.9kv+XQx2No0+Q2nUcsjQObxYo+bEOMC5agMGjmwEZRuSdAQoyKB0janFtuInW62dZuC2us+MX9fcfkQ0dziH3yu3zTthQzIrQ63.IHvX1PEIgkypW6RIOZzRS+91+FX3mAa.qAdBr16JM4QLESeSqkwyXovn3pFLGEzeF8et8GRmtu931GEL5b8ZwoRpdjQeE+7jfDVZrukrZUBFOBSJsWNSWjKHpcHfIK0gj.JQTnVjXSnnDNFWOk4es7p+0VGVo8oY7pIC06wC7Xhuh7BBrAMFHM1kyb8AEAk1od8aJx3qvHILHwDhKeEgIIo3mKnokXcRwkFFehklijEq0wQpufTVLXn5eCt2mDAeUdEfpVk1bqfWB0Jg7bCy0fEB8vy1gt9Hvf.TsXYNwGEl+7F6P3OynC7wCeHd8Q6YEgBMg2mn9diw9Y9uXBNF10TfKlaHFCtF+fg.7+ip0lSffpHw3ppYhHH1ZM80uCJnuJiJdlIKtJHP0Pv7byAc2uKgcIzNOU2EDNhRYuBZDai77G+08IrYZQdiockql7pSsj.MSEYFWvtFt03cAQRo8AtsKIiIKeaxoJI.31ZPtHhc8L9LXTS83E4yMB0uLBA9a7Je+HYALgjbYEi9tcp637Oh1I.+.

You "just" have to implement your filter algorithms and manage per channel integrators (ch arg)

@orange Yes I did, but I don't see where this path has been modified, the width is still the one from the panel with no reduction

p1 = gr_peakreduct.createPath([a[0], a[1], a[2], a[3]], [0,1,0,-1], 1.0);

[a[0], a[1], a[2], a[3]] this is equivalent to just write a

@DanH Ive used it on three and it compiles without issues!

@Christoph-Hart I'm also experiencing this multi-channel problem in the Hardcoded Fx.
Are there any plans to fix it?

@aaronventure Definitely a bug. But random.
I think the probability of doing this increases as the network becomes more complex.

I think that these computers will be a real pain in the ass in the not-too-distant future. Especially on the artificial intelligence side

I have the same issue. Framex block doesn't work.

But can all this structure be done with frame2 block?

@ustk I'll try it when I go to the computer, but normally this should change automatically.

@Christoph-Hart After completing elements such as knobs, buttons, sliders, combo box, panels..etc. +1 cent for the Draggable filter panel

@Christoph-Hart you can see in the video that the network itself gets saved to a file. This happens by default when you Save XML.

Doesn't a network (in this case UpDown) have to be a file in order to be importable into another network as a node?

Why wouldn't hitting Save XML also save the changes in the external/child network?

Can something be done about the UX of making changes there where a reload of the entire host network (either by ejecting it and loading it back or by reloading HISE) is necessary when any changes are made to the external/child network and Save XML is executed?

Any networks that also host the changed network should automatically reload when the project is saved.

@orange if possible you can also use a slider pack to reduce the number of parameters (if that is a sensible choice for your node).

But 40 should be fine, the node interface might look ugly as hell though…

@Christoph-Hart I think it should be fixed now, Thanks!

https://github.com/christophhart/HISE/commit/fbc02d78d4d33b809ce4b2c07e4b29a52e036480

@aaronventure Nice one cheers!!

@d-healey said in Warning: undefined parameter 0:

@orange said in Warning: undefined parameter 0:

undefined parameter 0

That's telling you that the first parameter that you are passing to a function -- the thing in () -- is undefined.

Makes sense :) But it wouldn't warn previously

@dannytaurus said in Warning: undefined parameter 0:

@orange Is there a line number shown with the error message?

Yes the code that I shared above is the dedicated line.

Sorry for the late reply. I just realized that I'm using a 3rd party node for my LUT test.
Here's the full project you guys can take a look:
https://hub.korneffaudio.com/index.php/s/xPcEwQicPH63Fkd

@orange Yeah I removed that because it was super glitchy.

@orange If low ratio is 1, high works independently for me:

highalone.gif

@orange In french, "GPT" sounds exactly like "I farted" (j'ai pété)... Maybe that's just the real sense of it all in the end, with a real philosophical thought behind...