Systematic RL training debugger. Four-phase diagnostic process for identifying and fixing RL training failures. Use when (1) reward not growing or plateaued, (2) gradient explosion or vanishing, (3) entropy collapsed or policy died, (4) Q-values diverging in SAC, (5) user says "debug training", "reward stuck", "why isn't it learning", (6) validate trainer before real training, (7) run probe environments, (8) user mentions "explained variance", "advantage stats", "action collapse", or any diagnostic metric, (9) pre-flight check before training, (10) reward function debugging, reward component analysis, reward SNR, (11) environment verification, observation sufficiency, action space issues, (12) task design assessment, curriculum necessity, exploration problems. Works with PPO, SAC, MM-RKHS, and any TorchRL-based trainer.

Albatross6790

Machine Learning

PINN Training Debugger

Systematic PINN training debugger. Four-phase diagnostic process for identifying and fixing physics-informed neural network training failures. Use when (1) loss not decreasing or stuck, (2) physics loss dominating or data loss dominating, (3) gradient pathology between loss components, (4) NaN in training, (5) user says "debug PINN", "loss stuck", "why isn't it learning", (6) validate PINN pipeline before real training, (7) run probe PDEs, (8) user mentions "loss ratio", "NTK eigenvalues", "ReLoBRaLo", "spectral bias", "Fourier features", "collocation", "residual", or any PINN diagnostic metric, (9) pre-flight check before PINN training, (10) physics residual debugging, per-component violation analysis, nondimensionalization quality, (11) convergence rate mismatch, causality violation, collocation starvation, (12) ansatz numerical issues, DD-PINN-specific debugging. Works with DD-PINN (train_pinn.py) and physics-regularized surrogates (train_regularized.py).

Albatross6790

Machine Learning

ML Training Pipeline Checklist

$38

Albatross6790

Machine Learning

Chemometrics Hybrid Modeling

Guide for combining mechanistic models with machine learning (hybrid modeling) in chemometrics and chemical engineering. Covers physics-informed ML, residual modeling, model augmentation, and constraint incorporation for improved predictions and interpretability.

AlbanOtt0

Machine Learning

Low-Float Catalyst Scanner

Pattern recognition framework for identifying low-float momentum opportunities using the ROLR case study (436% gain, January 2026) as reference. Use when asked to find stocks like ROLR, scan for low-float runners, identify gap-up momentum plays, find catalyst-driven small-caps, or evaluate pre-market gappers. Triggers on requests like "find low float stocks", "scan for momentum runners", "find stocks like ROLR", "pre-market gap scanners", "small cap momentum plays", "catalyst gap trading", or "help me find the next ROLR".

AkhilGurrapu0

Machine Learning

Mix Analyzer Module

This skill should be used when creating new audio analysis modules for Mix Analyzer, modifying existing analysis modules, or understanding the 17-module analysis pipeline architecture. Triggers on requests like "add a new analysis module", "create a loudness analyzer", "extend the frequency analysis", or "how do analysis modules work".

ahmedarafa9940

Machine Learning

Continuous Learning V2

Instinct-based learning system that observes sessions via hooks, creates atomic instincts with confidence scoring, and evolves them into skills/commands/agents. v2.1 adds project-scoped instincts to prevent cross-project contamination.

Ahmed2020SE0

Machine Learning

Orchestrator Prd Executor

Use when the user asks for a large PRD-driven change. This skill orchestrates atomic execution, context budgeting, and verification so smaller models can do most of the work.

AndMuruGvart0

Machine Learning

Client Pricing

Commercial pricing engine for freelance deliverables. Encodes the pricing hierarchy, tiered anchor-floor tables, negotiation decision tree, and rate collapse prevention.

andreagongrac-maker0

Machine Learning

Decision Journal

Unified decision lifecycle: Pre-decision logging, post-decision review, failure classification, and calibration tracking. Absorbs: post-mortem-engine.

andreagongrac-maker0

Machine Learning

Grpo Rl Training

Expert guidance for GRPO/RL fine-tuning with TRL for reasoning and task-specific model training

adm-humanerd0

Machine Learning

Unsloth

Expert guidance for fast fine-tuning with Unsloth - 2-5x faster training, 50-80% less memory, LoRA/QLoRA optimization

adm-humanerd0

Machine Learning

Co Scientist Ensemble Methods

Ensemble methods skill. Random forests, gradient boosting (XGBoost/LightGBM/CatBoost), stacking, blending, voting classifiers, and ensemble model interpretation. Use when working with random forests, gradient boosting (xgboost/lightgbm/catboost), stacking.

nahisaho0

Machine Learning

Test Engineer

Especialista em QA e testes automatizados. Garante cobertura, confiabilidade e previne regressões.

andviana230

Machine Learning

Muriel

Next-gen visual-production skill for LLMs — ten channels of tool-use recipes (raster, SVG, web, interactive, video, science, gaze, terminal, dimensions, style-guides), a two-tier brand-token schema with motion, anti-patterns per channel, a multi-constraint solver that enforces 8:1 contrast at render time, and a vision-model critique agent grounded in Tufte / Bertin / Gestalt. Use when the user needs any visual artifact for human eyes.

andyed0

Machine Learning

Continuous Learning V2

Instinct-based learning system that observes sessions via hooks, creates atomic instincts with confidence scoring, and evolves them into skills/commands/agents.

AndyHsuTW0

Machine Learning

Fine Tuning With Trl

Fine-tune LLMs using reinforcement learning with TRL - SFT for instruction tuning, DPO for preference alignment, PPO/GRPO for reward optimization, and reward model training. Use when need RLHF, align model with preferences, or train from human feedback. Works with HuggingFace Transformers.

adm-humanerd0

Machine Learning

Trainheroic Migration Prep

Prepare TrainHeroic export data for importing into Train's simplified database schema. Use when given a TrainHeroic export folder (especially training_data.csv) and asked to produce deterministic, import-ready files for workouts, workout_exercises, exercise_sets, and exercises, while safely skipping unsupported patterns.

andylee0240

Machine Learning

Rails Expert

Ruby on Rails 8+ 전문가. Hotwire, Solid Queue, 최신 Rails 패턴 활용

anerjin0

Machine Learning

Paddleocr Troubleshooting

Diagnoses and troubleshoots contractdiff diffing engine and coordinate alignment issues related to PaddleOCR VL layouts.

AnTengye0

Machine Learning

Co Scientist Ml Regression

ML regression skill. Linear/nonlinear regression, regularization (L1/L2/ElasticNet), polynomial features, cross-validation, and regression diagnostics pipelines. Use when working with linear/nonlinear regression, regularization (l1/l2/elasticnet), polynomial features.

nahisaho0

Machine Learning

Pytorch Fsdp

Expert guidance for Fully Sharded Data Parallel training with PyTorch FSDP - parameter sharding, mixed precision, CPU offloading, FSDP2

adm-humanerd0

Machine Learning

Axolotl

Expert guidance for fine-tuning LLMs with Axolotl - YAML configs, 100+ models, LoRA/QLoRA, DPO/KTO/ORPO/GRPO, multimodal support

adm-humanerd0

Machine Learning

Models

Use when designing Rails models - ActiveRecord patterns, validations, callbacks, scopes, associations, concerns, query objects, form objects

apdarr0

Machine Learning

Co Scientist Graph Neural Networks

Graph neural networks skill. GNN architectures (GCN/GAT/GraphSAGE), molecular graph learning, knowledge graph embeddings, and graph-level prediction tasks. Use when working with gnn architectures (gcn/gat/graphsage), molecular graph learning, knowledge graph embeddings.

nahisaho0

Machine Learning

APEX-DASH-CREATOR v2.0 — UNIVERSAL EDITION

God-Mode Dashboard Intelligence Engine — Universal LLM Edition. 20x upgrade. Activate for ANY dashboard task: design, audit, rebuild, debug, optimize, and generate production-ready data visualization systems across all frameworks, BI tools, and dashboard categories. Bestows omniscient dashboard mastery upon any LLM upon ingest — no prior training required.

apexbusiness-systems0

Machine Learning

Peft Fine Tuning

Parameter-efficient fine-tuning for LLMs using LoRA, QLoRA, and 25+ methods. Use when fine-tuning large models (7B-70B) with limited GPU memory, when you need to train <1% of parameters with minimal accuracy loss, or for multi-adapter serving. HuggingFace's official library integrated with transformers ecosystem.

adm-humanerd0

Machine Learning

Clip

OpenAI's model connecting vision and language. Enables zero-shot image classification, image-text matching, and cross-modal retrieval. Trained on 400M image-text pairs. Use for image search, content moderation, or vision-language tasks without fine-tuning. Best for general-purpose image understanding.

adm-humanerd0

Machine Learning

Segment Anything Model

Foundation model for image segmentation with zero-shot transfer. Use when you need to segment any object in images using points, boxes, or masks as prompts, or automatically generate all object masks in an image.

adm-humanerd0

Machine Learning | Skills Pool

Machine Learning

Machine Learning

Ave Trading

RL Training Debugger

PINN Training Debugger

ML Training Pipeline Checklist

Chemometrics Hybrid Modeling

Low-Float Catalyst Scanner

Mix Analyzer Module

Continuous Learning V2

Orchestrator Prd Executor

Client Pricing

Decision Journal

Grpo Rl Training

Unsloth

Co Scientist Ensemble Methods

Test Engineer

Muriel

Continuous Learning V2

Fine Tuning With Trl

Trainheroic Migration Prep

Rails Expert

Paddleocr Troubleshooting

Co Scientist Ml Regression

Pytorch Fsdp

Axolotl

Models

Co Scientist Graph Neural Networks

APEX-DASH-CREATOR v2.0 — UNIVERSAL EDITION

Peft Fine Tuning

Clip

Segment Anything Model