The Neural Amp Modeler (NAM), is a free (and open source) profiling guitar amplifier from Steven Atkinson. With Neural Amp Modeler, you can create algorithmic copies of your own guitar amplifiers, share them with others in the community, and enjoy unparalleled sound quality. This plugin is a major competitor for Kemper, Line 6 Helix, and Amplitube Tonex.
Any links for repositories of shared amps?
Edit: to download all the captures in one go. Click the green “Code” button at the top right, then “Download ZIP” at the bottom of that. That will package all the captures up into a ZIP archive and download it for you.
I see how a user could build his own IR file, but I don’t understand how a regular user could be able to build an amp « model »? I think I am missing something here… 
You don’t build an amp model using components or anything like that.
Let’s say you have a physical amp setup in your house, complete with mics and cabinets set the way you want. You then “train” the neural modeler using the training program they supply. I haven’t gone through this process myself, so don’t pretend to know the details, but that’s what they mean by “build a model.”
You can also build such a model using existing models. So let’s say I had a TH-U based rack that I really like the sound of. I could “train” NAM to that rackspace. Then I can share that “model” on the site and anyone else can load that and use it. Sort of like impulse responses in a sense.
Thanks for the clarification, so it’s a neural model associated to a learning phase, rather than a traditional analytical model. Well, why not… 
Warning: I’m talking here about Clean or Crunch type sounds.
After a few tries, big disappointment for me: in my case, not having an amp anymore (I switched to virtual for a few years) I was going to use NAM to simulate some of my favourite amps I used to own (Twin Reverb, Bogner, Fuchs ODS…).
Despite the number of models available and the IR libraries I have, I haven’t found anything that comes close to the level of the virtual amps I have (NeuralDSP, S-Gear, Heptode Amplosion, Nembrini, etc).
This is the major problem with simulations or amp learning, which are amazing on big hard and metal sounds (most of the demos, by the way) but which quickly become weak and not very credible on clean sounds.
I would like to have some feedback from those who are going to do a learning of their amps and cabs to know if the tracking of attack, dynamics and harmonics is also consistent in use.
After some discussions on the Audiofanzine forum, I got some recommendations that allowed me to find some very satisfying amp models that make me reconsider my thoughts about clean sounds.
I did not find a satisfactory Fender Twin model, but a Vox AC10 model gives me full satisfaction.
Quick test : Neural Amp Modeler - Vox AC10 - Clean
Thank you for sharing, every information and feedback is useful for other community members (and yet not registered users! 
).
I like the result, but I can’t find the ingredients for the recipe… 
Not sure how to download things from ToneX. Created an account, but no downloads, not even a shopping cart…
I’m sorry I can’t help you, I don’t know ToneX at all.
Ah. No Tonex included. That also helps, but what did you use?
I do not know what is a tonex.
Have a look at this video: at 0’44" for the FX chain and at 0’48"for the IR settings
I use the same settings except for the amp model: I use AC10 Clean instead of AC10 NewCorners.
A lot of amp models can be found here: NAM_models
See my post above.
I didn’t find this model in the nam models, so I thought you got it from tonex, but you didn’t clearly ( bad assumption on my side)
Sorry, I remember now: it comes from facebook but I do not have a facebook account.
My neighbour picked it up for me. You have to look for “AC10 - Dan Lombardi”.
And Ignite NadIR for the IR (included in the free Emissary bundle ) Ignite Amps
I will have to buy flowers for my wife, I guess (not a bad thought anyway): she’s on Facebook 
Thanks for explaining
ToneHunt website is beginning to compile a lot of NAM captures. e.g. AC-10
The plugin to load these captures is on the GitHub page.
