neural-amp-modeler

Neural network emulator for guitar amplifiers
Log | Files | Refs | README | LICENSE

colab.rst (3181B)


      1 Training in the cloud with Google Colab
      2 =======================================
      3 
      4 If you don't have a good computer for training ML models, you use Google Colab
      5 to train in the cloud using the pre-made Jupyter notebook at 
      6 `notebook.ipynb <https://github.com/sdatkinson/NAMTrainerColab/blob/main/notebook.ipynb>`_, 
      7 which is designed to be used with 
      8 `Google Colab <https://colab.research.google.com/>`_.
      9 
     10 Opening the notebook
     11 --------------------
     12 
     13 To open the notebook in Colab, follow 
     14 `this link <https://colab.research.google.com/github/sdatkinson/NAMTrainerColab/blob/main/notebook.ipynb>`_.
     15 
     16 .. note:: Most browsers work, but Firefox can be a bit temperamental. This isn't
     17     NAM's fault; Google Colab just prefers Chrome (unsurprisingly).
     18 
     19 You'll be met with a screen like this:
     20 
     21 .. image:: media/colab/welcome.png
     22 
     23 Reamping: Getting data for your model
     24 -------------------------------------
     25 
     26 In order to train, you're going to need data, which means you're going to need 
     27 an amp or a pedal you want to model, and you're going to need to have gear to 
     28 reamp with it. Start by downloading the standardized test signal here:
     29 
     30 .. image:: media/colab/get-input.png
     31     :scale: 20 %
     32 
     33 If you need help with reamping, others 
     34 `on YouTube <https://www.youtube.com/results?search_query=reamping+tutorial>`_
     35 have made high-quality tutorials.
     36 
     37 .. note:: You need to make sure that your exported file is the same length as 
     38     the input file. To help with this, the standardized input files are an 
     39     exact number of seconds long. If you drop them into a DAW session at 120 
     40     BPM, you can snap your guides to the beat and easily get the reamp of the 
     41     right length.
     42 
     43 However, if you want to skip reamping for your first model, you can download 
     44 these pre-made files:
     45 
     46 * `input.wav <https://drive.google.com/file/d/1KbaS4oXXNEuh2aCPLwKrPdf5KFOjda8G/view?usp=sharing>`_, 
     47   a standardized input file.
     48 * `output.wav <https://drive.google.com/file/d/1NrpQLBbCDHyu0RPsne4YcjIpi5-rEP6w/view?usp=sharing>`_, 
     49   a reamp of a high-gain tube head.
     50 
     51 To upload your data to Colab, click the Folder icon here:
     52 
     53 .. image:: media/colab/file-icon.png
     54     :scale: 50 %
     55 
     56 and either drag and drop the files into the panel or select them after clicking 
     57 the upload button.
     58 
     59 .. image:: media/colab/upload.png
     60 
     61 **Wait for the files to finish uploading before proceeding.** If you don't, then
     62 strange errors will happen.
     63 
     64 Training
     65 --------
     66 
     67 At this point, you can train your model with a single click: just click the Play
     68 button and everything will finish in about 10 minutes.
     69 
     70 .. image:: media/colab/1-click-train.png
     71 
     72 However, there are a lot of options below that you can use to tweak the training
     73 that are worth getting familiar with.
     74 
     75 TODO: explain the options.
     76 
     77 Downloading your model
     78 ----------------------
     79 
     80 Once training is done, you can download your model as a .nam file from the file 
     81 browser:
     82 
     83 .. image:: media/colab/download.png
     84     :scale: 20 %
     85 
     86 If you don't see it, you might have to refresh the file browser:
     87 
     88 .. image:: media/colab/refresh.png
     89     :scale: 20 %
     90 
     91 To use it, point 
     92 `the plugin <https://github.com/sdatkinson/NeuralAmpModelerPlugin>`_ at the file
     93 and you're good to go!