Machine Learning.
Focused GNCore specialist skills for machine learning tasks. Discover 23 documented skills, read their use cases, and copy a verified installation command.
Computer Vision
Build vision workflows with robust image handling, leakage-safe splits, and task-appropriate evaluation.
View skill and install command ↗02 / data-preparationML Data Preparation
Prepare reproducible model datasets while preventing leakage and preserving provenance.
View skill and install command ↗03 / deep-learningDeep Learning Engineering
Develop deep learning code with explicit tensor shapes, stable optimization, and reproducible evaluation.
View skill and install command ↗04 / evaluateEvaluate a Model
Evaluate model quality with leakage-safe splits, task-appropriate metrics, baselines, and error analysis.
View skill and install command ↗05 / experiment-trackingReproducible ML Experiments
Make model experiments traceable across code, data, configuration, environment, and artifacts.
View skill and install command ↗06 / feature-engineeringFeature Engineering
Build features that are valid at prediction time and reproducible across training and inference.
View skill and install command ↗07 / feature-engineering-practiceFeature Engineering Practice
Create leakage-safe, interpretable features that match the model's serving-time information.
View skill and install command ↗08 / fine-tuningModel Fine-Tuning
Plan or implement a parameter-efficient or full fine-tuning workflow with controlled data and cost.
View skill and install command ↗09 / hardware-mlEdge and Hardware-Constrained ML
Adapt inference to constrained hardware using measured latency, memory, and accuracy tradeoffs.
View skill and install command ↗10 / improveImprove a Machine Learning Model
Find evidence-based model improvements through error analysis and controlled experiments against a fixed baseline.
View skill and install command ↗11 / machine-learningMachine Learning Workflow
Develop a scoped machine learning feature, from data inspection and baseline selection through evaluation.
View skill and install command ↗12 / ml-experiment-trackingMachine Learning Experiment Tracking
Make model experiments comparable with consistent splits, metrics, configuration, and artifact provenance.
View skill and install command ↗13 / mlopsMLOps and Model Release
Package, register, and monitor models with reproducible artifacts and explicit release gates.
View skill and install command ↗14 / model-cardModel Documentation
Document a model's intended use, data provenance, evaluation, limitations, and deployment assumptions.
View skill and install command ↗15 / model-evaluationModel Evaluation
Evaluate model quality with valid splits, suitable metrics, uncertainty, and error analysis.
View skill and install command ↗16 / model-servingModel Serving and Inference APIs
Package model inference behind a reliable API with validation, resource limits, and observable behavior.
View skill and install command ↗17 / model-trainingModel Training
Set up or improve a reproducible model training run with clear configuration, checkpoints, and evaluation.
View skill and install command ↗18 / nlpNatural Language Processing
Build NLP pipelines that respect language, label quality, privacy, and task-specific evaluation.
View skill and install command ↗19 / preprocessPreprocess Machine Learning Data
Design a repeatable preprocessing pipeline that preserves schema, prevents leakage, and records transformations.
View skill and install command ↗20 / pytorch-trainingPyTorch Training
Implement PyTorch training loops with correct device, optimizer, checkpoint, and evaluation handling.
View skill and install command ↗21 / tensorflow-trainingTensorFlow and Keras Training
Build TensorFlow/Keras training workflows with explicit data, callbacks, and evaluation behavior.
View skill and install command ↗22 / time-seriesTime-Series Modeling
Analyze and forecast time-indexed data with temporal splits, leakage controls, and useful uncertainty.
View skill and install command ↗23 / trainTrain a Machine Learning Model
Build and run a reproducible model training workflow with a baseline, tracked configuration, and held-out evaluation.
View skill and install command ↗Choose first.
Then install.
Install only the skills relevant to your task and target. GNCore skills guide an AI assistant—they cannot perform work independently.
Read installation details ↗gncore install computer-vision --apps vscode-codex --scope workspace