nanoresearch-experiment
The nanoresearch-experiment skill generates a complete, runnable Python project structure from an experiment blueprint JSON file. It creates data loaders, model stubs, training loops with checkpointing, evaluation harnesses, ablation configurations, and a main entry point. Use this skill after the planning phase to transform high-level experiment designs into executable research code scaffolding.
git clone --depth 1 https://github.com/OpenRaiser/NanoResearch /tmp/nanoresearch-experiment && cp -r /tmp/nanoresearch-experiment/skills/nanoresearch-experiment ~/.claude/skills/nanoresearch-experimentSKILL.md
# Experiment Skill ## Purpose Take the experiment blueprint and produce a runnable Python code skeleton that implements the proposed method, baselines, training loops, evaluation harness, and ablation configurations. ## Tools Required None. This skill operates entirely through LLM code generation based on the experiment blueprint. ## Input - `experiment_blueprint`: Path to `papers/experiment_blueprint.json` produced by the planning skill ## Process 1. Parse the experiment blueprint for datasets, baselines, metrics, and ablation groups 2. Generate the project directory structure (data loaders, models, training, evaluation, configs) 3. Produce data loading and preprocessing code for each specified dataset 4. Implement model architecture stubs for the proposed method and each baseline 5. Generate training loop with logging, checkpointing, and early stopping 6. Implement the evaluation harness computing all specified metrics 7. Create configuration files for each ablation group 8. Add a main entry point that accepts a config and runs the full train-evaluate pipeline ## Output Produces `experiments/` directory containing: - `data/`: Data loading and preprocessing modules - `models/`: Model architecture implementations (proposed method and baselines) - `training/`: Training loop and optimization utilities - `evaluation/`: Metric computation and result aggregation - `configs/`: YAML configuration files for each experiment and ablation variant - `run.py`: Main entry point for launching experiments - `requirements.txt`: Python dependencies
Generates publication-quality figures for ML papers from research context. Given a paper section or description, extracts system components and relationships to generate architecture diagrams via Gemini. Given experiment results or data, auto-selects chart type and generates data-driven figures via matplotlib/seaborn. Use when creating any figure for a conference paper.
Orchestrates end-to-end autonomous AI research projects using a two-loop architecture. The inner loop runs rapid experiment iterations with clear optimization targets. The outer loop synthesizes results, identifies patterns, and steers research direction. Routes to domain-specific skills for execution, supports continuous agent operation via Claude Code /loop and OpenClaw heartbeat, and produces research presentations and papers. Use when starting a research project, running autonomous experiments, or managing a multi-hypothesis research effort.
Guides researchers through structured ideation frameworks to discover high-impact research directions. Use when exploring new problem spaces, pivoting between projects, or seeking novel angles on existing work.
Applies cognitive science frameworks for creative thinking to CS and AI research ideation. Use when seeking genuinely novel research directions by leveraging combinatorial creativity, analogical reasoning, constraint manipulation, and other empirically grounded creative strategies.
Evaluates LLMs across 60+ academic benchmarks (MMLU, HumanEval, GSM8K, TruthfulQA, HellaSwag). Use when benchmarking model quality, comparing models, reporting academic results, or tracking training progress. Industry standard used by EleutherAI, HuggingFace, and major labs. Supports HuggingFace, vLLM, APIs.
Simplest distributed training API. 4 lines to add distributed support to any PyTorch script. Unified API for DeepSpeed/FSDP/Megatron/DDP. Automatic device placement, mixed precision (FP16/BF16/FP8). Interactive config, single launch command. HuggingFace ecosystem standard.
Write publication-ready ML/AI/Systems papers for NeurIPS, ICML, ICLR, ACL, AAAI, COLM, OSDI, NSDI, ASPLOS, SOSP. Use when drafting papers from research repos, structuring arguments, verifying citations, or preparing camera-ready submissions. Includes LaTeX templates, reviewer guidelines, and citation verification workflows.
Battle-tested PyTorch training recipes for all domains — LLMs, vision, diffusion, medical imaging, protein/drug discovery, spatial omics, genomics. Covers training loops, optimizer selection (AdamW, Muon), LR scheduling, mixed precision, debugging, and systematic experimentation. Use when training or fine-tuning neural networks, debugging loss spikes or OOM, choosing architectures, or optimizing GPU throughput.