I did the install without nVidia GPU on WIndows 11 and I get the error "OSError: [WinError 126] can't find this module." the module is fbgemm.dll and it is where is should be...any idea why it cannot be loaded? thanks!
Hard to say without seeing. One thing that trips people up often is that you have to close the plots that pop up before it continues. If that's not it, then you can always try asking for help in the community.
@@RunawayThumbtack another problem has poped out，i use my own wave file in 48KHZ\24Bit\mono by ADOBE AUDITION. But when i begin to train, the terminal tell m e this file cannot be recognized as any known version, can NAM Trainer use my own file? And where is the community link or website? Best regards :)
any numbers in comparison to a mid-range PC with mid-range GPU? how many epochs oer minute? is this STILL working with later generations, M2/M3 or even M4 now? cheers
A mid-range nVIDIA GPU will usually be better than Apple Silicon. Accelerator support (GPUs, Apple Silicon, etc) is based on the underlying ML framework, PyTorch. Since PyTorch is perhaps the most popular deep learning framework in the world, you can be sure they're working to support as much as possible :)
NAM's GPU support is limited to what PyTorch supports. Currently, AMD GPUs are only supported by PyTorch on Linux. So you can use your AMD GPU if you install on Linux.
ERROR: pip's dependency resolver does not currently take into account all the packages that are installed. This behaviour is the source of the following dependency conflicts. anaconda-cloud-auth 0.1.4 requires pydantic<2.0, but you have pydantic 2.7.1 which is incompatible.
Can you please make bug reports over at the GitHub repo? github.com/sdatkinson/neural-amp-modeler/issues This will help me make sure I don't lose it :)
I get the following error message when trying to enter the prompt in step 3: "pip : The term 'pip' is not recognized as the name of a cmdlet, function, script file, or operable program. Check the spelling of the name, or if a path was included, verify that the path is correct and try again. At line:1 char:2"
thats not the hard part tbh. all you gotta do is import and route the impulse file (input) out of the daw (i recommend first unused output, usually 3) out of output x (i used 3) into a reamp box, from the reamp box into an amp, mic up the cab and plug the amp into the first unused input of the interface (for me thats 1) and back into the DAW make a new track with input x (i used 1) and then enable record arm and hit record. once the end was reached trim it to the same location as the impulse file (input) and then export it.
I appreciate that but I really still don’t understand …I know how to reamp so that concept I understand …I also think I understand that the new track /file I create needs to match the one that’s in the program ?….am I recording my amp or just the file through my amp ?
@@JonRayRunawayVealMusic "am I recording my amp or just the file through my amp ?" You reamp the source audio file provided by NAM through your (captured) chain (amp, pedal, etc).
Tremolo could be 2 short waves and produce this effect. However If this means lengthy captures work now reverb, delays and other time effects can be profiled ?
It is a long term -short term neural net architecture so it can capture longer term evolution in principle.. I guess it depends on the training material and on the network memory 'depth' but it is not surprising at least on the principle... But this is clearly the first time I see that demonstrated in a (convincing) example !
This is for training a model, not playing. So you need to provide a pair of files to use as data for the model. For the input, grab the file from here: neural-amp-modeler.readthedocs.io/en/latest/tutorials/gui.html Then, reamp your gear for the output file. Hope this helps :)
Which website? neuralampmodeler.com? Anaconda is legit 🙂 It's not technically absolutely necessary to install, but it's certainly more convenient (especially on Windows!). If you're unfamiliar with it, a little bit of Googling around should give you an Idea of how common it is, especially with machine learning and data science 🙂
Lost …looks really great . I’ve only been able to use what others have created….for that thank you . I would need a much more hands on tutorial …it’s me …just seems too complex for me personally
Can't get this to work. After installing and trying to do step #3, I try and open the application _conda and the window briefly opens then closes again quickly. I'm stuck! How do I open the Anaconda prompt?
For other people as frustrated as I was with these instructions, there is apparently a step missing. Following these instructions, Anaconda would always tell me that "nam is not recognized as an internal or external command". Before typing "nam" you need to activate the environment by tying "conda activate nam". I don't think it will work without this step and I spent HOURS troubleshooting this with several folks who are professional devs. Really frustrating to have incomplete documentation like this, perhaps something has changed since this video was made?
Sorry for the confusion. Typing "conda activate nam" should only be necessary if you _also_ typed that before doing the installation steps in this video (that would have caused you to install nam into that environment). Notice the "(base)" in the command prompt. That's the default conda environment; I'm installing to that for simplicity since that's what you'll get dropped into automatically when you open the Anaconda prompt. The steps in the video are still correct as of my writing this and should work if you follow them as I did them. (Don't mix and match with others' instructions!)
Input and output are the signal you used to reamp and what you recorded from your gear, respectively. Train destination is where the .nam files will get written. Hope that helps! If you still have questions about the process, the online training video may answer some questions since it uses the same sort of reamp pair.
thank you. i got the program running but after i click on train it says "Plotting the delay for manual inspection" in the terminal and i see a window with a graph but nothing else happens. im on mac m1. should this work?
Wow, so easy while not understanding a single thing :-). Now I mangaged to open nam trainer, I will try my first model this weekend I hope... THANK YOU!!!
Hi, I‘ve been using Nam for many weeks now and I absolutely love it! But now I‘ve changed my Operating System to Win7 because of better Audio performance and somehow the Nam plugin doesn‘t work… any fix?
@@josefplager this is a limitation of the iPlug2 framework that's used by this particular plugin. Since NAM is open-source, there are limitless ways to fix it--but you need to be able to write code.
