Agent skill

Deepmd Train

Train DeePMD-kit machine learning potentials. An agent skill from Hello-QM/catgo-LRG.

AGPL-3.0Auto-check passedAI & LLM Engineering

Install Deepmd Train

skills CLI

$ npx skills add Hello-QM/catgo-LRG --skill deepmd-train -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI

$ gh skill install Hello-QM/catgo-LRG deepmd-train --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy

$ git clone --depth 1 https://github.com/Hello-QM/catgo-LRG.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.claude/skills/deepmd-train .claude/skills/deepmd-train && rm -rf skills-src

Use ~/.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/

Paste into Claude Code

Install the "deepmd-train" agent skill from https://github.com/Hello-QM/catgo-LRG/tree/main/.claude/skills/deepmd-train into .claude/skills/deepmd-train/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "deepmd-train", 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

$skill-installer install https://github.com/Hello-QM/catgo-LRG/tree/main/.claude/skills/deepmd-train

Type 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.

skills CLI

$ npx skills add Hello-QM/catgo-LRG --skill deepmd-train -a codex

Project install goes to .agents/skills/; add -g for ~/.codex/skills/.

GitHub CLI

$ gh skill install Hello-QM/catgo-LRG deepmd-train --agent codex

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/Hello-QM/catgo-LRG.git skills-src && mkdir -p .agents/skills && cp -r skills-src/.claude/skills/deepmd-train .agents/skills/deepmd-train && rm -rf skills-src

Use ~/.agents/skills/ instead of .agents/skills for a personal install.

Codex skills documentation · loads skills from .agents/skills/

Paste into Codex

Install the "deepmd-train" agent skill from https://github.com/Hello-QM/catgo-LRG/tree/main/.claude/skills/deepmd-train into .agents/skills/deepmd-train/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "deepmd-train", 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.

skills CLI

$ npx skills add Hello-QM/catgo-LRG --skill deepmd-train -a cursor

Project install goes to .agents/skills/; add -g for ~/.cursor/skills/.

GitHub CLI

$ gh skill install Hello-QM/catgo-LRG deepmd-train --agent cursor

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/Hello-QM/catgo-LRG.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/.claude/skills/deepmd-train .cursor/skills/deepmd-train && rm -rf skills-src

Use ~/.cursor/skills/ instead of .cursor/skills for a personal install.

Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/

Paste into Cursor

Install the "deepmd-train" agent skill from https://github.com/Hello-QM/catgo-LRG/tree/main/.claude/skills/deepmd-train into .cursor/skills/deepmd-train/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "deepmd-train", 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

$ gemini skills install https://github.com/Hello-QM/catgo-LRG.git --path .claude/skills/deepmd-train

--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.

skills CLI

$ npx skills add Hello-QM/catgo-LRG --skill deepmd-train -a gemini-cli

Project install goes to .agents/skills/; add -g for ~/.gemini/skills/.

GitHub CLI

$ gh skill install Hello-QM/catgo-LRG deepmd-train --agent gemini-cli

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/Hello-QM/catgo-LRG.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/.claude/skills/deepmd-train .gemini/skills/deepmd-train && rm -rf skills-src

Use ~/.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/

Paste into Gemini CLI

Install the "deepmd-train" agent skill from https://github.com/Hello-QM/catgo-LRG/tree/main/.claude/skills/deepmd-train into .gemini/skills/deepmd-train/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "deepmd-train", 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.

GitHub CLI

$ gh skill install Hello-QM/catgo-LRG deepmd-train

Installs 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).

skills CLI

$ npx skills add Hello-QM/catgo-LRG --skill deepmd-train -a github-copilot

Project install goes to .agents/skills/; add -g for ~/.copilot/skills/.

Manual copy

$ git clone --depth 1 https://github.com/Hello-QM/catgo-LRG.git skills-src && mkdir -p .github/skills && cp -r skills-src/.claude/skills/deepmd-train .github/skills/deepmd-train && rm -rf skills-src

Use ~/.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/

Paste into GitHub Copilot

Install the "deepmd-train" agent skill from https://github.com/Hello-QM/catgo-LRG/tree/main/.claude/skills/deepmd-train into .github/skills/deepmd-train/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "deepmd-train", 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.

skills CLI

$ npx skills add Hello-QM/catgo-LRG --skill deepmd-train -a opencode

OpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.

GitHub CLI

$ gh skill install Hello-QM/catgo-LRG deepmd-train --agent opencode

Project scope by default (.agents/skills/); add --scope user for a personal install.

Manual copy

$ git clone --depth 1 https://github.com/Hello-QM/catgo-LRG.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/.claude/skills/deepmd-train .opencode/skills/deepmd-train && rm -rf skills-src

Use ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.

OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/

Paste into OpenCode

Install the "deepmd-train" agent skill from https://github.com/Hello-QM/catgo-LRG/tree/main/.claude/skills/deepmd-train into .opencode/skills/deepmd-train/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "deepmd-train", 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.

Facts

Skill name: deepmd-train
GitHub stars: 205
Used in: 1 other repo
Token cost: ~1.1k tokens
SKILL.md length: 283 words
Files: 1
Skills in repo: 75
Repo updated: Sep 22, 2026
First seen: Oct 7, 2026
Licence: AGPL-3.0

At a glance

Train DeePMD-kit machine learning potentials. An agent skill from Hello-QM/catgo-LRG.

Works in 3 steps: Create workflow → Add training task → Add freeze step (convert to production…
Tasks that involve Fine-tuning
SKILL.md covers When to Use, Prerequisites, Data Directory Structure and Workflow Steps, plus 5 more sections
Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md

What it does

Deepmd Train is an agent skill from Hello-QM/catgo-LRG. Train DeePMD-kit machine learning potentials. Covers DPA-3 (recommended), see2a (legacy), and fine-tuning from pretrained models.

Its SKILL.md is about 1.1k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts. Compatibility notes: Requires deepmd-kit = 3.0, GPU with CUDA. Training data in dpdata format.

It sits in AI & LLM Engineering, covering Fine-tuning and Machine learning. The repository describes itself as: AI-driven workbench for computational materials science — interactive 3D structure viewer, natural-language CatBot assistant, visual DAG workflow engine, HPC job submission… The licence is AGPL-3.0.

When your agent uses it

Tasks that involve Fine-tuning
Tasks that involve Machine learning

Example prompts

“/deepmd-train”

Requirements

Compatibility (from SKILL.md): Requires deepmd-kit >= 3.0, GPU with CUDA. Training data in dpdata format.

Workflow steps

3 steps, taken from the step headings in SKILL.md.

Create workflow
Add training task
Add freeze step (convert to production model)

What it can do on your machine

Read from SKILL.md and the folder at commit fd6291b. It shows what the files ask for, not the result of running them.

Tool permissions
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.
Runs code
No scripts in the folder and no shell commands in SKILL.md (its code samples are json and bash).
From the folder's file list and the shell code blocks in SKILL.md.
Network
No URLs in SKILL.md.
From URLs in SKILL.md, links to its own repository left out.
Credentials
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Compatibility
Requires deepmd-kit >= 3.0, GPU with CUDA. Training data in dpdata format.
From compatibility in the SKILL.md frontmatter.

Context cost

Deepmd Train loads about 1.1k tokens when it runs. Until then it costs about 36 tokens; SKILL.md has 283 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it: ~36
When it runs · the whole SKILL.md, loaded when a task matches: ~1.1k

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.

Safety

Auto-check passed

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); files beside SKILL.md are not scanned.

SKILL.md

The full file from Hello-QM/catgo-LRG at commit fd6291b, republished under its AGPL-3.0 licence (© Hello-QM). 283 words, ~1,146 tokens.

Download SKILL.mdSave it as .claude/skills/deepmd-train/SKILL.md (or your agent's skills folder).

name: deepmd-train
description: Train DeePMD-kit machine learning potentials. Covers DPA-3 (recommended), se_e2_a (legacy), and fine-tuning from pretrained models.
compatibility: Requires deepmd-kit >= 3.0, GPU with CUDA. Training data in dpdata format.
catalog-hidden: true

DeePMD Training

When to Use

User wants to train a machine learning interatomic potential
User has DFT data (energies, forces, stresses) and wants a DP model
User wants to fine-tune a pretrained DPA-3 model

Prerequisites

Training data in DeePMD format (use data/dpdata/SKILL.md to convert from VASP/QE)
GPU node available on HPC
deepmd-kit installed (dp --version)

Data Directory Structure

data/
├── train/
│   ├── set.000/
│   │   ├── coord.npy       # Atomic coordinates (natoms*3,)
│   │   ├── energy.npy      # Total energy (1,)
│   │   ├── force.npy       # Forces (natoms*3,)
│   │   ├── box.npy         # Cell vectors (9,)
│   │   └── virial.npy      # Stress tensor (optional, 9,)
│   └── type.raw            # Atom type indices
└── valid/
    └── set.000/
        └── ...

Workflow Steps

1. Create workflow

catgo_workflow_engine(action="create", params={"name": "DeePMD DPA-3 training"})

2. Add training task

catgo_workflow_engine(action="add_task", params={
  "workflow_id": "wf_xxx",
  "task_type": "shell",
  "name": "dp_train",
  "command": "dp train input.json 2>&1 | tee train.log",
  "input_files": {
    "input.json": "<training config>"
  },
  "system_name": "dp_model"
})

3. Add freeze step (convert to production model)

catgo_workflow_engine(action="add_task", params={
  "workflow_id": "wf_xxx",
  "task_type": "shell",
  "name": "dp_freeze",
  "command": "dp freeze -o frozen_model.pb",
  "depends_on": ["dp_train"],
  "system_name": "dp_model"
})

DPA-3 Training Config (Recommended)

json

{
  "model": {
    "type_map": ["Ti", "O"],
    "descriptor": {
      "type": "dpa3",
      "rcut": 6.0,
      "rcut_smth": 0.5,
      "sel": "auto",
      "neuron": [25, 50, 100],
      "n_interaction": 3,
      "n_head": 4
    },
    "fitting_net": {
      "type": "ener",
      "neuron": [240, 240, 240]
    }
  },
  "training": {
    "training_data": {
      "systems": ["./data/train"],
      "batch_size": "auto"
    },
    "validation_data": {
      "systems": ["./data/valid"],
      "batch_size": "auto"
    },
    "numb_steps": 1000000,
    "disp_freq": 1000,
    "save_freq": 10000
  },
  "learning_rate": {
    "type": "exp",
    "start_lr": 1e-3,
    "stop_lr": 1e-8,
    "decay_steps": 5000
  },
  "loss": {
    "type": "ener",
    "start_pref_e": 0.02,
    "limit_pref_e": 1.0,
    "start_pref_f": 1000,
    "limit_pref_f": 1.0,
    "start_pref_v": 0.0,
    "limit_pref_v": 0.0
  }
}

se_e2_a Training Config (Legacy)

Replace the descriptor block:

json

"descriptor": {
  "type": "se_e2_a",
  "rcut": 6.0,
  "rcut_smth": 0.5,
  "sel": [46, 92],
  "neuron": [25, 50, 100],
  "axis_neuron": 16
}

sel must list the max neighbor count per element type. Use dp neighbor-stat to determine.

Fine-Tuning from Pretrained Model

bash

# Download pretrained model (e.g., from AIS Square)
# Then fine-tune:
dp train input.json --finetune pretrained.pb 2>&1 | tee finetune.log

Reduce numb_steps to 100000-200000 and start_lr to 1e-4 for fine-tuning.

Parameter Guidance

Parameter	Typical value	Notes
rcut	6.0-9.0 A	Interaction cutoff; larger = more accurate but slower
sel	auto or [N1, N2]	Max neighbors per type; use `dp neighbor-stat`
numb_steps	500K-2M	More data needs more steps
start_lr	1e-3	Learning rate; reduce for fine-tuning
batch_size	auto	Let DeePMD choose based on system size
start_pref_f	1000	Force weight starts high, decays to limit_pref_f

Common Pitfalls

Insufficient training data — need at least 1000-5000 frames for a reliable model. More diverse configs = better.
sel too small — if max neighbors exceeds sel, training crashes. Always run dp neighbor-stat first.
Overfitting — monitor validation loss. If train loss drops but valid loss plateaus, stop training or add more data.
Missing virial — if training data has no stress tensor, set start_pref_v = 0, limit_pref_v = 0.
Mixed element sets — type_map must be consistent across all training systems and inference.
Forgetting to freeze — the checkpoint directory is not the production model. Run dp freeze to create a portable .pb file.

Files

Just SKILL.md in .claude/skills/deepmd-train of Hello-QM/catgo-LRG.

Open the folder on GitHubat commit fd6291b

Used in 1 other repository

We found 3 copies of this SKILL.md (exact, near-identical or edited) in other folders, from 1 other GitHub owner. This page covers the copy in Hello-QM/catgo-LRG, which our catalogue first saw on October 7, 2026.

majiayu000/claude-skill-registryskills/ai-ml/train skills/ai-ml/deepmd-train

Compare with similar skills

Deepmd Train 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.

Deepmd Train compared with similar skills
Skill	Stars	Used in	Tokens	Auto-check	Licence	Repo updated
Deepmd Train this skillHello-QM/catgo-LRG	205	1 repos	~1.1k	Automated safety check: Pass	AGPL-3.0	16 days ago
Unimoljinzhezenggroup/computational-chemistry-agent-skills	148	1 repos	~1.5k	Automated safety check: Pass	LGPL-3.0-or-later	3 days ago
Adapting Transfer Learning Modelsjeremylongshore/tons-of-skills-marketplace	2.8k	—	~1.1k	Automated safety check: Pass	MIT	yesterday
ML Fairchem Finetunelearningmatter-mit/AtomisticSkills	176	—	~1.7k	Automated safety check: Pass	MIT	yesterday
ML Mace Finetunelearningmatter-mit/AtomisticSkills	176	—	~2.6k	Automated safety check: Pass	MIT	yesterday
ML Matgl Finetunelearningmatter-mit/AtomisticSkills	176	—	~1.4k	Automated safety check: Pass	MIT	yesterday

Similar skills

Unimol
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…
148 GitHub starsUsed in 1 repo~1.5k tokens
AI & LLM EngineeringAuto-check passed
Adapting Transfer Learning Models
jeremylongshore/tons-of-skills-marketplace
Build this skill automates the adaptation of pre-trained machine learning models using transfer learning techniques.
2.8k GitHub stars~1.1k tokensUpdated yesterday
AI & LLM EngineeringAuto-check passed
ML Fairchem Finetune
learningmatter-mit/AtomisticSkills
Fine-tune Fairchem machine learning interatomic potentials (UMA, ESEN) on custom datasets.
176 GitHub stars~1.7k tokensUpdated yesterday
AI & LLM EngineeringAuto-check passed
ML Mace Finetune
learningmatter-mit/AtomisticSkills
Fine-tune MACE machine learning interatomic potentials on custom datasets.
176 GitHub stars~2.6k tokensUpdated yesterday
AI & LLM EngineeringAuto-check passed
ML Matgl Finetune
learningmatter-mit/AtomisticSkills
Fine-tune MatGL machine learning interatomic potentials on custom datasets.
176 GitHub stars~1.4k tokensUpdated yesterday
AI & LLM EngineeringAuto-check passed
ML Research Lab
AnastasiyaW/codex-claude-code-config
Machine-learning research loop for dataset curation, fine-tuning, evaluation, inference deployment, experiment tracking, and model explainability.
154 GitHub stars~794 tokensUpdated yesterday
AI & LLM EngineeringAuto-check passed

Questions about Deepmd Train

What does Deepmd Train do?

Train DeePMD-kit machine learning potentials. An agent skill from Hello-QM/catgo-LRG. Deepmd Train is an agent skill from Hello-QM/catgo-LRG. Train DeePMD-kit machine learning potentials.

When should I use Deepmd Train?

Deepmd Train fits situations like: tasks that involve Fine-tuning; tasks that involve Machine learning.

How do I install Deepmd Train in Claude Code?

Run `npx skills add Hello-QM/catgo-LRG --skill deepmd-train -a claude-code`. Or copy the skill folder (.claude/skills/deepmd-train in Hello-QM/catgo-LRG) into .claude/skills/deepmd-train in your project. Claude Code loads it when a task matches its description.

How do I install Deepmd Train in Codex?

Run `npx skills add Hello-QM/catgo-LRG --skill deepmd-train -a codex`. Or copy the skill folder (.claude/skills/deepmd-train in Hello-QM/catgo-LRG) into .agents/skills/deepmd-train in your project. Codex loads it when a task matches its description.

Can I use Deepmd Train in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add Hello-QM/catgo-LRG --skill deepmd-train -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/deepmd-train, .gemini/skills/deepmd-train, .github/skills/deepmd-train and .opencode/skills/deepmd-train in your project.

What does Deepmd Train need to run?

SKILL.md names no scripts, command-line tools or credentials: Deepmd Train is instructions for the agent only. Compatibility (from SKILL.md): Requires deepmd-kit >= 3.0, GPU with CUDA. Training data in dpdata format. .

Does Deepmd Train access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Deepmd Train safe to install?

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. Review the folder before installing.

What licence does Deepmd Train use?

Deepmd Train is published under the AGPL-3.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Deepmd Train use?

About 1.1k tokens (SKILL.md is roughly 4.6k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Deepmd Train?

Skills that share tags, products or a category with Deepmd Train: Unimol (jinzhezenggroup/computational-chemistry-agent-skills, 148 stars), Adapting Transfer Learning Models (jeremylongshore/tons-of-skills-marketplace, 2.8k stars), ML Fairchem Finetune (learningmatter-mit/AtomisticSkills, 176 stars) and ML Mace Finetune (learningmatter-mit/AtomisticSkills, 176 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Deepmd Train?

Hello-QM (a GitHub user) maintains it in Hello-QM/catgo-LRG, which has 205 GitHub stars. The repository holds 75 skills in this directory. The repository was last updated on September 22, 2026.

Source: Hello-QM/catgo-LRG on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.

Install Deepmd Train

Facts

At a glance

What it does

When your agent uses it

Example prompts

Requirements

Workflow steps

What it can do on your machine

Tool permissions

Runs code

Network

Credentials

Compatibility

Context cost

Safety

SKILL.md

DeePMD Training

When to Use

Prerequisites

Data Directory Structure

Workflow Steps

1. Create workflow

2. Add training task

3. Add freeze step (convert to production model)

DPA-3 Training Config (Recommended)

se_e2_a Training Config (Legacy)

Fine-Tuning from Pretrained Model

Parameter Guidance

Common Pitfalls

Files

Used in 1 other repository

Compare with similar skills

Similar skills

More from Hello-QM/catgo-LRG

Related

Questions about Deepmd Train