kermt-infer

Execute KERMT model inference workflows on SMILES chemical datasets.

413|62|Updated Jun 23, 2026

Search Tags:#cheminformatics #inference #smiles #molecular-docking #deep-learning #kermt

Ready to run

npx skills add https://github.com/NVIDIA-BioNeMo/bionemo-agent-toolkit --skill kermt-infer

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: kermt-infer
Source: https://github.com/NVIDIA-BioNeMo/bionemo-agent-toolkit/tree/main/plugins/bionemo-agent-toolkit/skills/kermt-infer
Command: npx skills add https://github.com/NVIDIA-BioNeMo/bionemo-agent-toolkit --skill kermt-infer

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires docker, nvidia-container-toolkit, rdkit, python, and includes scripts (resource) components.

What problem does it solve?

This skill automates the complex inference pipeline for KERMT models, removing the manual burden of validating checkpoints, preparing SMILES data, and managing containerized execution environments.

Core Features & Use Cases

Automated Validation: Ensures checkpoints have the required task FFN heads and validates input CSV formats before execution.
Data Preparation: Automatically cleans SMILES strings and generates RDKit 2D features required for model inference.
Use Case: A researcher needs to predict binding properties for a large library of candidate molecules; this skill handles the entire end-to-end workflow from raw CSV to structured prediction output.

Quick Start

Use the kermt-infer skill to run predictions on molecules.csv using the finetuned checkpoint located at /path/to/model.ckpt.

Frequently Asked Questions about kermt-infer

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I run high-throughput molecular property predictions from SMILES strings?▼

To run high-throughput molecular property predictions from SMILES strings, you can use an automated inference pipeline that cleans raw CSV inputs, generates RDKit 2D features, and executes containerized deep learning models on chemical datasets.

How does automated SMILES data preparation work for deep learning molecular docking?▼

Automated SMILES data preparation for molecular docking works by validating input CSV formats, cleaning chemical strings, and generating the required RDKit 2D features needed for deep learning model inference.

Do I need a CUDA-capable GPU and Docker to run molecular inference workflows?▼

Yes, you need a CUDA-capable environment with Docker and the NVIDIA container toolkit to execute high-performance containerized model inference for molecular property prediction.

Can I use a finetuned checkpoint with missing task FFN heads for molecular predictions?▼

No, you cannot use a finetuned checkpoint with missing task FFN heads for molecular predictions because automated validation ensures checkpoints have the required task FFN heads before execution proceeds.

What is the best way to validate input CSV formats for cheminformatics model inference?▼

The best way to validate input CSV formats for cheminformatics model inference is using an automated validation step that checks checkpoint structures and SMILES data formats before executing containerized workflows.

Why does my KERMT inference workflow fail to generate RDKit 2D features?▼

Your KERMT inference workflow may fail to generate RDKit 2D features if the input SMILES strings are not properly cleaned or if the required Python RDKit dependency is missing from your environment.