Fine-tune MACE machine learning interatomic potentials on custom datasets.
$ npx skills add learningmatter-mit/AtomisticSkills --skill ml-mace-finetune -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install learningmatter-mit/AtomisticSkills ml-mace-finetune --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/learningmatter-mit/AtomisticSkills.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/ml-mace-finetune .claude/skills/ml-mace-finetune && rm -rf skills-srcUse ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.
Claude Code skills documentation · loads skills from .claude/skills/
Install the "ml-mace-finetune" agent skill from https://github.com/learningmatter-mit/AtomisticSkills/tree/main/skills/ml-mace-finetune into .claude/skills/ml-mace-finetune/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "ml-mace-finetune", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/learningmatter-mit/AtomisticSkills/tree/main/skills/ml-mace-finetuneType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add learningmatter-mit/AtomisticSkills --skill ml-mace-finetune -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install learningmatter-mit/AtomisticSkills ml-mace-finetune --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/learningmatter-mit/AtomisticSkills.git skills-src && mkdir -p .agents/skills && cp -r skills-src/skills/ml-mace-finetune .agents/skills/ml-mace-finetune && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "ml-mace-finetune" agent skill from https://github.com/learningmatter-mit/AtomisticSkills/tree/main/skills/ml-mace-finetune into .agents/skills/ml-mace-finetune/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "ml-mace-finetune", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add learningmatter-mit/AtomisticSkills --skill ml-mace-finetune -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install learningmatter-mit/AtomisticSkills ml-mace-finetune --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/learningmatter-mit/AtomisticSkills.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/skills/ml-mace-finetune .cursor/skills/ml-mace-finetune && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "ml-mace-finetune" agent skill from https://github.com/learningmatter-mit/AtomisticSkills/tree/main/skills/ml-mace-finetune into .cursor/skills/ml-mace-finetune/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "ml-mace-finetune", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/learningmatter-mit/AtomisticSkills.git --path skills/ml-mace-finetune--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add learningmatter-mit/AtomisticSkills --skill ml-mace-finetune -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install learningmatter-mit/AtomisticSkills ml-mace-finetune --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/learningmatter-mit/AtomisticSkills.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/skills/ml-mace-finetune .gemini/skills/ml-mace-finetune && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "ml-mace-finetune" agent skill from https://github.com/learningmatter-mit/AtomisticSkills/tree/main/skills/ml-mace-finetune into .gemini/skills/ml-mace-finetune/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "ml-mace-finetune", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install learningmatter-mit/AtomisticSkills ml-mace-finetuneInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add learningmatter-mit/AtomisticSkills --skill ml-mace-finetune -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/learningmatter-mit/AtomisticSkills.git skills-src && mkdir -p .github/skills && cp -r skills-src/skills/ml-mace-finetune .github/skills/ml-mace-finetune && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "ml-mace-finetune" agent skill from https://github.com/learningmatter-mit/AtomisticSkills/tree/main/skills/ml-mace-finetune into .github/skills/ml-mace-finetune/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "ml-mace-finetune", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add learningmatter-mit/AtomisticSkills --skill ml-mace-finetune -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install learningmatter-mit/AtomisticSkills ml-mace-finetune --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/learningmatter-mit/AtomisticSkills.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/skills/ml-mace-finetune .opencode/skills/ml-mace-finetune && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "ml-mace-finetune" agent skill from https://github.com/learningmatter-mit/AtomisticSkills/tree/main/skills/ml-mace-finetune into .opencode/skills/ml-mace-finetune/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "ml-mace-finetune", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
ml-mace-finetuneFine-tune MACE machine learning interatomic potentials on custom datasets.
ML Mace Finetune is an agent skill from learningmatter-mit/AtomisticSkills. Fine-tune MACE machine learning interatomic potentials on custom datasets.
Its SKILL.md is about 2.6k tokens, which your agent loads only when the skill is triggered. The skill folder holds 12 other files, including scripts (for example `examples/mace-wbm-finetune/README.md`, `examples/mace-wbm-finetune/finetune_config.yaml` and `examples/mace-wbm-finetune/training_history.json`).
It sits in AI & LLM Engineering, covering Fine-tuning and Machine learning. The repository describes itself as: Integrating AtomisticSkills into Agentic IDEs (Cursor, Claude Code, Codex, Google Antigravity, Hermes Agent, etc). The licence is MIT.
8 steps, taken from the first numbered list in SKILL.md.
Read from SKILL.md and the folder at commit 7f2d86d. It shows what the files ask for, not the result of running them.
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Ships 3 files in scripts/ (Python), which the agent can run.
From the folder's file list and the shell code blocks in SKILL.md.
Links to these hosts (documentation or services it may open):
github.comFrom URLs in SKILL.md, links to its own repository left out.
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
ML Mace Finetune loads about 2.6k tokens when it runs. Until then it costs about 23 tokens; SKILL.md has 1,051 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); the scripts in this folder are not scanned.
The full file from learningmatter-mit/AtomisticSkills at commit 7f2d86d, republished under its MIT licence (© learningmatter-mit). 1,051 words, ~2,574 tokens.
.claude/skills/ml-mace-finetune/SKILL.md (or your agent's skills folder). This skill also uses 9 other files; get the full folder from GitHub.To evaluate and improve the accuracy of a foundation MACE potential for a specific chemical system or physical property using the provided Python fine-tuning script and data-augmentation.
/benchmark-finetuning workflow for details.scripts/prepare_mace_data.py to convert JSON structures to .xyz data files.scripts/generate_mace_config.py using the .xyz data to produce finetune_config.yaml.mace_run_train --config /path/to/finetune_config.yaml to begin fine-tuning natively on the GPU.register_model tool to register the newly fine-tuned model checkpoint into the local registry so future research tasks can discover and reuse it.MACE fine-tuning is divided into a data preparation step, a configuration generation step, and a standard native training run. The script scripts/prepare_mace_data.py generates .xyz files, and scripts/generate_mace_config.py converts arguments into a fully-formed finetune_config.yaml configuration compatible with the MACE default parser.
| Key | Type | Default | Description |
|---|---|---|---|
--data | str | (Required) | Path to JSON file containing ASE/pymatgen structure dictionaries |
--output-dir | str | ./fine_tuning_data | Directory to save the converted .xyz data |
--val-split | float | 0.1 | Fraction of data to set aside for validation |
--seed | int | 42 | Random seed for validation splitting |
--vasp-stress-conversion | flag | - | If set, multiplies stress values by -1/160.2x to convert VASP raw kB to eV/ų |
| Key | Type | Default | Description |
|---|---|---|---|
--train-file | str | (Required) | Path to the converted train.xyz data |
--valid-file | str | None | Path to the valid.xyz data |
--model | str | MACE-MP-small | Base model name or path to a checkpoint |
--epochs | int | 100 | Number of training epochs |
--lr | float | 0.01 | Peak learning rate for training |
--batch-size | int | 2 | Training batch size |
--output-dir | str | ./fine_tuning | Directory to save the fine-tuned model and logs |
--device | str | cuda | Target compute device (cuda or cpu) |
[!NOTE] If you have created a dedicated research directory for your current workflow (e.g. using the
create_research_dirtool), you should set the--output-dirargument to a folder within that active research directory to keep all artifacts and models organized.
| Key | Type | Default | Choices | Description |
|---|---|---|---|---|
--freeze-backbone | flag | N/A | Add flag | Freeze backbone (interaction blocks); only readout heads are trained. |
--reinit-head | flag | N/A | Add flag | Re-initialize readout weights. When absent (default), pre-trained readout is preserved. |
--multiheads | flag | N/A | Add flag | Enable multi-head fine-tuning (adds new head while keeping existing ones). |
[!IMPORTANT] The following parameters govern Optimizer, Regularization, and Scheduling. They are NOT exposed via the
prepare_mace_data.pyCLI. Note thatprepare_mace_data.pyexposes--energy-weight,--forces-weight(default 10.0), and--stress-weightwhich inject cleanly into the initial YAML. For all other properties below, you must manually append the keys to the generatedfinetune_config.yamlfile prior to runningmace_run_train.
| Key | Type | Default | Choices / Range | Description |
|---|---|---|---|---|
optimizer | str | "adam" | "adam", "adamw", "schedulefree" | Optimizer type. |
weight_decay | float | 5e-7 | ≥0 | L2 weight decay. |
amsgrad | bool | True | True, False | Use AMSGrad variant of Adam. |
clip_grad | float | 10.0 | >0 or None | Maximum gradient norm for clipping. Set to None to disable. |
ema | bool | True | True, False | Enable exponential moving average of model weights. |
ema_decay | float | 0.99 | 0–1 | EMA decay rate. Higher = more smoothing. |
| Key | Type | Default | Choices / Range | Description |
|---|---|---|---|---|
scheduler | str | "ReduceLROnPlateau" | "ReduceLROnPlateau", "ExponentialLR" | LR scheduler type. |
lr_factor | float | 0.8 | 0–1 | Factor by which LR is reduced on plateau (for ReduceLROnPlateau). |
scheduler_patience | int | 50 | ≥1 | Epochs without improvement before reducing LR (for ReduceLROnPlateau). |
lr_scheduler_gamma | float | 0.9993 | 0–1 | Per-epoch multiplicative decay factor (for ExponentialLR). |
ReduceLROnPlateau (default): Monitors validation loss. When loss stops improving for scheduler_patience epochs, LR is multiplied by lr_factor.
ExponentialLR: LR decays every epoch by lr_scheduler_gamma. At epoch n: LR = learning_rate × lr_scheduler_gamma^n.
| Key | Type | Default | Choices / Range | Description |
|---|---|---|---|---|
patience | int | 2048 | ≥1 | Stop training after this many epochs without improvement. |
[!NOTE] Default
patience=2048effectively disables early stopping. Set lower (e.g., 100–200) if you want to stop early.
| Key | Type | Default | Choices | Description |
|---|---|---|---|---|
loss | str | "weighted" | "weighted", "universal", etc. | Loss function type. "universal" is auto-set by data script when stress data is detected. |
energy_weight | float | 1.0 | ≥0 | Weight for energy loss (exposed via CLI). |
forces_weight | float | 10.0 | ≥0 | Weight for forces loss (exposed via CLI). |
stress_weight | float | 1.0 | ≥0 | Weight for stress loss (exposed via CLI). |
compute_forces | bool | True | True, False | Include forces in training. |
compute_stress | bool | False | True, False | Include stress in training (auto-enabled by data script). |
[!WARNING] Stress Units: MACE expects stress in
eV/ų. Raw VASP stress obtained directly via some JSON files may be in kilo-Bar (kB), which is ~160x larger and will cause catastrophic training divergence. The Atomate2 MCP tool handles this conversion automatically whenconvert_units=True. However, if your JSON labels contain rawkBstress, you MUST pass the--vasp-stress-conversionflag toscripts/prepare_mace_data.pyto automatically scale them by-1/160.2x. For more details on unit standardization, see @[skills/general-property-units/SKILL.md].
[!WARNING] Learning rate sensitivity for MACE-OMAT: The official MACE docs recommend
lr=0.01for MACE-MP-0, but MACE-OMAT-0-small requireslr=1e-4to avoid divergence. Higher values (1e-3, 0.01) cause catastrophic forgetting even with frozen backbone + EMA.
[!IMPORTANT] The generated
finetune_config.yamlmaps exactly tomace_run_trainarguments. You can open and manually modify the YAML file before runningmace_run_trainto inject any advanced parameter.
Usage:
# 1. Prepare Data
${CLAUDE_SKILL_DIR}/../../venv/run mlip python ${CLAUDE_SKILL_DIR}/scripts/prepare_mace_data.py \
--data /path/to/training_data.json \
--output-dir ./mace_finetuned_data
# 2. Generate Configuration
${CLAUDE_SKILL_DIR}/../../venv/run mlip python ${CLAUDE_SKILL_DIR}/scripts/generate_mace_config.py \
--train-file ./mace_finetuned_data/train.xyz \
--valid-file ./mace_finetuned_data/valid.xyz \
--model MACE-OMAT-0-small \
--epochs 10 \
--lr 1e-4 \
--batch-size 2 \
--freeze-backbone \
--output-dir ./mace_finetuned
# 3. Run Training
${CLAUDE_SKILL_DIR}/../../venv/run mlip mace_run_train --config ./mace_finetuned/finetune_config.yaml
# 4. Extract Training Logs (Optional, to create standard training_history.json)
${CLAUDE_SKILL_DIR}/../../venv/run cpu python ${CLAUDE_SKILL_DIR}/scripts/extract_mace_logs.py \
--results-dir ./mace_finetuned/resultsSee the mace-wbm-finetune directory for a complete, runnable example of 10-epoch fine-tuning on high-energy crystal structures (WBM dataset), including exact usage of the --vasp-stress-conversion flag and diagnostic output artifacts.
freeze_backbone=True is strongly recommended.training_history.json files use meV units: energy MAE in meV/atom, force MAE in meV/Å, stress MAE in meV/ų.Author: Bowen Deng Contact: GitHub @learningmatter-mit
© learningmatter-mit, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
SKILL.md and 9 other files (scripts) in skills/ml-mace-finetune of learningmatter-mit/AtomisticSkills.
Open the folder on GitHubat commit 7f2d86d
ML Mace Finetune next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| ML Mace Finetune this skilllearningmatter-mit/AtomisticSkills | 175 | — | ~2.6k | Automated safety check: Pass | MIT | |
| Unimoljinzhezenggroup/computational-chemistry-agent-skills | 148 | 1 repos | ~1.5k | Automated safety check: Pass | LGPL-3.0-or-later | |
| Deepmd TrainHello-QM/catgo-LRG | 205 | 1 repos | ~1.1k | Automated safety check: Pass | AGPL-3.0 | |
| ML Research LabAnastasiyaW/codex-claude-code-config | 154 | — | ~794 | Automated safety check: Pass | MIT | |
| RuView Model Trainingruvnet/RuView | 97k | — | ~1.3k | Automated safety check: Notes | MIT | |
| ML Experiment IterationLeeroo-AI/superml | 195 | — | ~4.8k | Automated safety check: Pass | Apache-2.0 |
jinzhezenggroup/computational-chemistry-agent-skills
A standardized CLI wrapper for Uni-Mol molecular ML workflows that handles representation extraction (embeddings), model training (regression/classification), and property prediction with built-in…
Hello-QM/catgo-LRG
Train DeePMD-kit machine learning potentials. An agent skill from Hello-QM/catgo-LRG.
AnastasiyaW/codex-claude-code-config
Machine-learning research loop for dataset curation, fine-tuning, evaluation, inference deployment, experiment tracking, and model explainability.
ruvnet/RuView
Trains and evaluates several WiFi-signal-based pose and sensing models, from unsupervised pose estimation to domain adaptation and publishing.
Leeroo-AI/superml
Produces ranked, evidence-grounded next steps when an ML experiment has stalled, drawing on a Leeroopedia knowledge base or on fetched docs and issues.
wshobson/agents
Fine-tune vision-language models (VLMs) with supervised learning on image+text data.
learningmatter-mit/AtomisticSkills
Define a docking search box (center coordinates + box dimensions in Angstroms) from a co-crystal ligand, binding-site residues, or a saved JSON specification.
learningmatter-mit/AtomisticSkills
Build a solvated, charge-neutralized protein-ligand complex for OpenMM molecular dynamics simulation.
learningmatter-mit/AtomisticSkills
Identify and rank ligandable pockets on a protein structure or model using geometry (fpocket) or an ML predictor (P2Rank).
learningmatter-mit/AtomisticSkills
Calculate homolytic and heterolytic bond dissociation energies (BDEs) for all single bonds in a molecule using MLIPs with RDKit fragmentation.
learningmatter-mit/AtomisticSkills
Generate molecular conformers with RDKit ETKDG, relax with MLIPs, and rank by energy with Boltzmann weighting.
learningmatter-mit/AtomisticSkills
Query multiple MOF databases (QMOF via MPContribs; ARC-MOF DB7/Majumdar et al.
Categories
Fine-tune MACE machine learning interatomic potentials on custom datasets. ML Mace Finetune is an agent skill from learningmatter-mit/AtomisticSkills. Fine-tune MACE machine learning interatomic potentials on custom datasets.
ML Mace Finetune fits situations like: tasks that involve Fine-tuning; tasks that involve Machine learning.
Run `npx skills add learningmatter-mit/AtomisticSkills --skill ml-mace-finetune -a claude-code`. Or copy the skill folder (skills/ml-mace-finetune in learningmatter-mit/AtomisticSkills) into .claude/skills/ml-mace-finetune in your project. Claude Code loads it when a task matches its description.
Run `npx skills add learningmatter-mit/AtomisticSkills --skill ml-mace-finetune -a codex`. Or copy the skill folder (skills/ml-mace-finetune in learningmatter-mit/AtomisticSkills) into .agents/skills/ml-mace-finetune in your project. Codex loads it when a task matches its description.
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add learningmatter-mit/AtomisticSkills --skill ml-mace-finetune -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/ml-mace-finetune, .gemini/skills/ml-mace-finetune, .github/skills/ml-mace-finetune and .opencode/skills/ml-mace-finetune in your project.
Going by SKILL.md and its folder, ML Mace Finetune needs Python for the scripts in its folder. Our summary lists: Python 3.
SKILL.md names 1 domain. As links in the text: github.com. This is read from the text; nothing was executed.
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. The check reads SKILL.md only: the scripts in the folder are not scanned, so read them before running anything.
ML Mace Finetune is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.
About 2.6k tokens (SKILL.md is roughly 10k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.
Skills that share tags, products or a category with ML Mace Finetune: Unimol (jinzhezenggroup/computational-chemistry-agent-skills, 148 stars), Deepmd Train (Hello-QM/catgo-LRG, 205 stars), ML Research Lab (AnastasiyaW/codex-claude-code-config, 154 stars) and RuView Model Training (ruvnet/RuView, 97k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
learningmatter-mit (a GitHub organization) maintains it in learningmatter-mit/AtomisticSkills, which has 175 GitHub stars. The repository holds 129 skills in this directory. The repository was last updated on October 6, 2026.
Source: learningmatter-mit/AtomisticSkills on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.
