Neural Amp Modeler
Yesterday, HeadRush released Firmware 5.1 for its Prime, Core, and Flex Prime units. This update introduces the ability to play Neural Amp Modeler's A2 and A1 architectures. How it works Here's how I got my HeadRush Prime running a model of my Fender Deluxe Reverb. I'm running Firmware 5.1.0. Get a model on the HeadRush Option 1: From my computer via USB I have a model of my EVH 5150III on my…
Today, I'm releasing Architecture "A2", NAM's new standard neural network recipe for modeling guitar and bass amps. Check out this YouTube video for a walkthrough: If you want a pre-made plugin that plays A2, you can check out Gateway over on the Users page. To train your own A2 models, TONE3000 now supports training A2 models, as does my open-source Google Colab notebook or the local GUI-based…
Published: April 30, 2026 NAM impact: NAM’s training code depends on PyTorch Lightning. Users who installed the affected Lightning versions should take action. Overview A supply-chain security incident has affected certain versions of PyTorch Lightning distributed via PyPI. Lightning AI has published an official advisory identifying versions 2.6.2 and 2.6.3 as compromised. These versions should be…
We're one step closer to A2 today, and I've released an update to NeuralAmpModelerCore, version 0.5.0: https://github.com/sdatkinson/NeuralAmpModelerCore/releases/tag/v0.5.0 Builders interested in how A2 works should aim to support core version 0.5.0 or later. Work isn't entirely compelte yet-- while 0.5.0 will be able to process A2 models correctly , optimized implementations are in development…
Progress on Architecture A2 has been coming along well, and I'm pretty happy with the results that we've gotten out of Stage 4 of the work (optimizing A2). With that, development will enter the fifth and final stage soon: listening tests . If you haven't already, please sign up to participate by filling out this Google Form . Please also share it with people you know--the more responses we get,…
This post is part of a series on Architecture A2 . I've been working on designing NAM's new default architecture, "A2". In this blog post, I wanted to give stakeholders some insight into what I've seen so far in Stage 4 ("Optimize A2", from the original blog post linked above). The work isn't done yet--but I do want to share some information and how it's influencing the way I'm thinking. If you're…
Update 2026-02-12: The test models have been updated to "v4" to fix some identified issues. Please re-download the latest if you haven't begun testing. If you've already sent me results, don't worry about it--I'll note the discrepancies and account for them. The core repo has also been updated to v0.4.0.rc3 Progress is coming along on work to develop Architecture A2 , NAM's update to the "standard…
In order to assist with Stage 2 of the Architecture A2's development , I've (finally!) created documentation for the real-time library, and I've integrated v0.4.0 into NeuralAmpModelerPlugin. Core library documentation I've completed a first pass at documentation for the library. You can access it online here: https://neuralampmodelercore.readthedocs.io/en/latest/ Aside from the usual API…
Update 2026-01-29: v0.4.0.rc2 is released with some updates. I've put up a release candidate for NeuralAmpModelerCore v0.4.0. You can access it here: https://github.com/sdatkinson/NeuralAmpModelerCore/releases/tag/v0.4.0.rc2 This is a release candidate and is hot off the press. If you encounter any bugs, please either create a bug report or get in contact with me. This release marks the completion…
Blog posts about A2: NeuralAmpModelerCore v0.4.0.rc2 is released (2026-01-21) NeuralAmpModelerCore documentation and demo integrations are live (2026-01-21) Test Models are available (2026-01-29) Sign up for listening tests (2026-04-13) In quarter 1 of 2026, I plan to develop and release a new default architecture for snapshot NAM models. I'm going to name this architecture " A2 ". Here's the…
I'm excited to share a paper I've recently completed: "Slimmable NAM: Slimmable NAM: Neural Amp Models with adjustable runtime computational cost" The paper was accepted for the 2025 NeurIPS workshop on Artificial Intelligence for Music: Where Creativity Meets Computation ; I'll be presenting it in 2 weeks in San Diego. In the meantime, anyone can access it here: https://arxiv.org/abs/2511.07470 .…
I recently released a new version of the Python package for training models. As usual, the newest version is automatically used when...
See also: this blog post on "NAM-to-X" The latest in the ecosystem of NAM-integrating products is NUX's Amp Academy Stomp, a...
In the last few months, we've seen an explosion of products from different companies that have benefited from the open-source nature of...
I recently released a new version of the Python package for training models. As usual, the newest version is automatically used when...
This January at the NAMM show, Valeton shared the GP-5 , a mini-format pedal they had been working on that included all of the staple...
Superior Drive is a macOS/iOS standalone/AU plugin developed by Artera DSP that gives users a parametric Neural Amp Model of an...
I'm excited to share that today, ML Sound Lab released Amped Volcano , a high-gain guitar plugin featuring a new modeling technology...
When I started NAM in 2018, the project was meant as a sort of "crossword puzzle" challenge for myself. I'd been picking up experience in...
After a year in the making, I’m very excited to share PreFire , a new plugin by SubMission Audio , which features parametric Neural Amp...