HISE Logo Forum
    • Categories
    • Register
    • Login

    8 Times more CPU consumption on Aida-X Neural Models

    Scheduled Pinned Locked Moved General Questions
    48 Posts 14 Posters 3.3k Views
    Loading More Posts
    • Oldest to Newest
    • Newest to Oldest
    • Most Votes
    Reply
    • Reply as topic
    Log in to reply
    This topic has been deleted. Only users with topic management privileges can see it.
    • DabDabD
      DabDab @orange
      last edited by

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

      I can load the Aida-X neural model in HISE

      Can you pleas make a guide video or post how to use neural network or model in HISE ? i am very excited to learn the new technology but didn't get proper resources. I created a thread on this forum but no luck.

      Bollywood Music Producer and Trance Producer.

      griffinboyG orangeO 2 Replies Last reply Reply Quote 0
      • griffinboyG
        griffinboy @DabDab
        last edited by

        @DabDab

        I've done it with tensorflow using google Colab. I can help you get started if you just need to get your foot in the door.

        DabDabD 1 Reply Last reply Reply Quote 1
        • DabDabD
          DabDab @griffinboy
          last edited by DabDab

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

          google Colab

          Oh.. nice. I am at the door. Ting Tong. 😀 Let me know what are the required steps. because I never tried google Colab and all that you have written.

          Bollywood Music Producer and Trance Producer.

          1 Reply Last reply Reply Quote 0
          • orangeO
            orange @DabDab
            last edited by orange

            @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.

            develop Branch / XCode 13.1
            macOS Monterey / M1 Max

            C DabDabD 2 Replies Last reply Reply Quote 2
            • C
              ccbl @orange
              last edited by

              Something else coming back to the performance difference. I vaguely remember saying that there were optimisations pre-compiled when it came to inferencing certain architecture sizes. Given that Aida for example has a pretty specific pipeline that people use on collab, maybe they optimised that specific number of Layers and Hidden Size?

              Christoph HartC 1 Reply Last reply Reply Quote 0
              • Christoph HartC
                Christoph Hart @ccbl
                last edited by

                @ccbl yes but none of that should cause an 8x performance boost (more like 20% or so). I just need to profile it and find out where it's spending its time.

                C 2 Replies Last reply Reply Quote 1
                • C
                  ccbl @Christoph Hart
                  last edited by

                  @Christoph-Hart cool. Well like I said, if you want any kind of models for testing, let me know. I can do either NAM-wavenet or LSTM, of any size.

                  1 Reply Last reply Reply Quote 0
                  • C
                    ccbl @Christoph Hart
                    last edited by

                    @Christoph-Hart One thing, in the Neural Example in the docs it says this "It requires HISE to be built with the RTNeural framework to enable real time inferencing..."

                    I don't remember doing this explicitly when I built HISE, I just built it the standard way and it all works. I'm assuming this is just no longer a requirement? Otherwise could it explain the performance penalty?

                    1 Reply Last reply Reply Quote 0
                    • DabDabD
                      DabDab @orange
                      last edited by

                      @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.

                      Bollywood Music Producer and Trance Producer.

                      1 Reply Last reply Reply Quote 0
                      • S
                        scottmire
                        last edited by

                        @orange I am probably doing something wrong, but when I try to make this work, HISE crashes. Here is my process:

                        • I copy the NN code from the Sine synth example into the interface onInit script in my project
                        • I replace everything in the obj declaration with the .json from one of your AIDA-X captures
                        • I add a scriptnode math.neural node in FX
                        • I select the NN obj in the dropdown
                        • Hise crashes.

                        Any thoughts?

                        1 Reply Last reply Reply Quote 0
                        • JulesVJ
                          JulesV
                          last edited by JulesV

                          @Christoph-Hart Has there been a fix for this?

                          Christoph HartC 1 Reply Last reply Reply Quote 0
                          • Christoph HartC
                            Christoph Hart @JulesV
                            last edited by

                            @JulesV I pushed the preliminary work I had on my mobile rig but it‘s not usable as it is. I‘ll need to do a few more performance tests and cleanup.

                            JulesVJ A orangeO FortuneF 4 Replies Last reply Reply Quote 2
                            • JulesVJ
                              JulesV @Christoph Hart
                              last edited by

                              @Christoph-Hart Thank you.

                              I see you also added ONNX. I'm not sure if we can open NAM models in ONNX, but it would be great to be able to load and run NAM models with performance.

                              1 Reply Last reply Reply Quote 0
                              • A
                                aaronventure @Christoph Hart
                                last edited by

                                @Christoph-Hart 3417c3ca-b4cc-42d4-a91f-77d28f66fb4c-image.png

                                peak developer humor

                                Christoph HartC 1 Reply Last reply Reply Quote 1
                                • Christoph HartC
                                  Christoph Hart @aaronventure
                                  last edited by

                                  @aaronventure haha I would love to be as funny as this, but in this case my old keyboard with the defect d key pranked me one last time... but damn that's a solid joke.

                                  1 Reply Last reply Reply Quote 1
                                  • orangeO
                                    orange @Christoph Hart
                                    last edited by

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

                                    @JulesV I pushed the preliminary work I had on my mobile rig but it‘s not usable as it is. I‘ll need to do a few more performance tests and cleanup.

                                    Has this issue been resolved? I still observe the same situation in the current commit.

                                    develop Branch / XCode 13.1
                                    macOS Monterey / M1 Max

                                    1 Reply Last reply Reply Quote 0
                                    • FortuneF
                                      Fortune @Christoph Hart
                                      last edited by Fortune

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

                                      @JulesV I pushed the preliminary work I had on my mobile rig but it‘s not usable as it is. I‘ll need to do a few more performance tests and cleanup.

                                      Almost 6 months passed. The problem still persists. Why is the neural node still unusable after all this time?

                                      You need to solve this problem as soon as possible, my friend.

                                      Christoph HartC 1 Reply Last reply Reply Quote -1
                                      • Christoph HartC
                                        Christoph Hart @Fortune
                                        last edited by

                                        @Fortune just don't use it then.

                                        FortuneF 1 Reply Last reply Reply Quote -1
                                        • FortuneF
                                          Fortune @Christoph Hart
                                          last edited by Fortune

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

                                          @Fortune just don't use it then.

                                          Wow, what an amazing answer you gave as the creator of HISE :)))

                                          It has nothing to do with me bro, the question is this:

                                          Will the Developers be able to use Neural networks with HISE or not?

                                          Will your HISE be compatible with the future or not? That is the question.

                                          It is obvious from your answer that you have failed to solve the source of the problem. But by acting like this, all you will do is alienate people and ruin your potential future businesses, nothing else.

                                          LindonL 1 Reply Last reply Reply Quote 0
                                          • LindonL
                                            Lindon @Fortune
                                            last edited by

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

                                            It has nothing to do with me bro,

                                            Then I guess its not a problem.

                                            HISE is an open source project - you cannot expect any given piece of functionality to get added and to be supported because you want it.

                                            However you are not stuck with the whims of @Christoph-Hart , if there's something you want and Christoph has not built it or not completed it to your satisfaction - then you can go build it yourself.

                                            HISE Development for hire.
                                            www.channelrobot.com

                                            Christoph HartC 1 Reply Last reply Reply Quote 3
                                            • First post
                                              Last post

                                            15

                                            Online

                                            1.7k

                                            Users

                                            11.9k

                                            Topics

                                            103.3k

                                            Posts