Skillget

oh-my-mlip

agent
Security Audit
Warn
Health Warn
  • License — License: MIT
  • Description — Repository has a description
  • Active repo — Last push 0 days ago
  • Low visibility — Only 6 GitHub stars
Code Pass
  • Code scan — Scanned 12 files during light audit, no dangerous patterns found
Permissions Pass
  • Permissions — No dangerous permissions requested

No AI report is available for this listing yet.

SUMMARY

Take the pain out of installing and running machine-learning interatomic potentials: every major MLIP in one registry, set up, verified and run by your LLM.

README.md

oh-my-mlip

oh-my-mlip — machine learning interatomic potentials

CI (GPU-free)
Docs
License

Every major MLIP in one place, run by your LLM. oh-my-mlip gathers 20
machine-learning interatomic-potential frameworks (32 model variants) behind
one registry: each installs into its own conda env from a pinned recipe, and
every one is installed, used, benchmarked, fine-tuned and distilled the same
way. Ask Claude Code or Codex, and it does the whole job for you. It drives the
real upstream frameworks and never reimplements a model.

Documentation: https://oh-my-mlip.org/
· Troubleshooting and help · Issues

Quick start

Claude Code — add the plugin, then ask "install MACE and check it runs on my GPU":

/plugin marketplace add JinukMoon/oh-my-mlip

/plugin install oh-my-mlip@oh-my-mlip

Codex or another coding agent — tell it:

Clone https://github.com/JinukMoon/oh-my-mlip, read its AGENTS.md and the
documentation at https://oh-my-mlip.org/, and use it for the
MLIP work I ask for.

From the terminal, without an agent:

git clone https://github.com/JinukMoon/oh-my-mlip.git && cd oh-my-mlip
source env.sh
./install.sh MACE                                  # MACE in its own conda env
python scripts/setup_verify.py MACE-MPA-0 --json   # energy + forces on your GPU

Gated models (UMA, eSEN) need a Hugging Face login first — see Gated models.

What it does
Install every framework in its own env from a pinned recipe, with exact replay from lock files
Use a model the exact interpreter and calculator lines for your own scripts, or one call across models
Benchmark CatBench adsorption benchmarks across models, including your own VASP calculations
Fine-tuning one command drives each framework's own trainer, config and dataset format
Distillation any hub model into an NN-MTP student that runs in LAMMPS

Supported MLIPs

Framework Models
SevenNet SevenNet-MF-OMPA, SevenNet-Omni
MACE MACE-MPA-0, MACE-MH-1-OMAT, MACE-MH-1-OC20
NequIP NequIP-OAM-XL, NequIP-OAM-L
Allegro Allegro-OAM-L
Nequix Nequix-MP-1
DeePMD DPA-3.1-3M-FT
ORB ORB-v3
GRACE GRACE-2L-OAM
MatterSim MatterSim-v1-5M
CHGNet CHGNet-v0.3.0
AlphaNet AlphaNet-v1-OMA
Eqnorm Eqnorm-MPtrj
fairchemv1 eSEN-30M-OAM
EquiformerV3 EqV3-OMatMPtrjSalex
UMA UMA-m-1p1-OC20, UMA-m-1p1-OMAT, UMA-s-1p1-OC20, UMA-s-1p1-OMAT, UMA-s-1p2-OC20, UMA-s-1p2-OC22, UMA-s-1p2-OC25, UMA-s-1p2-OMAT
PET PET-OAM-XL
EquFlash EquFlashV2, EquFlash-v1
MatRIS MatRIS-10M-OAM
DPA4 DPA-4.0.1-pro-MPtrj
TACE TACE-OAM-L

Weights, gating, licenses and upstream repositories: Supported models.

Contributing

New models and new tools are welcome. CONTRIBUTING.md covers how
to add a model to the registry, how to add a tool or workflow on top of the hub, and
the GPU proof a pull request needs.

License

MIT (LICENSE). Frameworks and weights keep their upstream licenses; the
distillation engine onthefly-distill
is a separate GPL-2.0 project.

Reviews (0)

No results found