For the love of God, don't use Feynman lectures to learn physics. That's something you read after you know physics, for relaxation and conceptual stuff. Resnick & Halliday is a much better freshman/sophomore book.

Susskind's "theoretical minimum" is actually pretty good.

http://theoreticalminimum.com/courses

Fowler gives a pretty conventional undergraduate physics curriculum (adding Feynman in there somehow). If it were me: learn the math tools first. I assume you know linear algebra; learn differential equations. From there, go straight to higher level books. There's very little difference in undergraduate vs graduate quantum mechanics and E&M other than the math is slightly more sophisticated in grad school. Might as well do it right. Messiah for QM and Jackson for E&M. Classical mechanics, the tradition is to learn Lagrangian mechanics in high level undergrad and Hamiltonian in grad school. There's no real reason to do it in this order, and a decent reason (understanding Quantum) to do it in reverse order. Amusingly, the math is cleaner in Hamiltonian mechanics, but you may find yourself unable to do some simple problems you can do with Newtonian physics; so this will be a weird working backward thing. Stat Mech, I think you should just read Reif; skip Ma or whatever they use in grad school now.

FWIIW I know/knew people who did this: started grad school without having done any undergrad courses in physics. I think skipping a lot of the introductory stuff, and visiting it later is actually better.

The rest of it can be done with the same machinery you learned in QM, E&M, Mechanics and Stat Mech. Max leverage if you had to pick one: probably classical mechanics for a DL guy, E&M for general knowledge of tools.

I'd suggest not actually trying to simulate physical systems on a computer: you probably stare at computers too much anyway.