VectorSpaceLab/AREX-Skill - Page 3

provider-connectors

AgentScope provider, embedding, formatter, and TTS configuration workflows.

rag-memory

AgentScope retrieval, vector-store, and long-term memory workflows.

service-platform

AgentScope FastAPI service, storage, hub, channel, MCP, and deployment workflows.

workspace-sandboxes

AgentScope local and sandboxed workspace backend workflows.

agibot-x1-train

Guide agents through AgiBot X1 humanoid reinforcement-learning training, checkpoint playback, policy export, and MuJoCo sim2sim workflows with verified configuration contracts and explicit Isaac Gym limits.

export

Route AgiBot X1 DH checkpoint-to-JIT export, JIT-to-ONNX conversion, artifact preflight, validation, and backend-aware failure recovery.

playback

Prepare, locate, and safely route interactive Isaac Gym playback of trained AgiBot X1 locomotion checkpoints, including viewer and Logitech F710 control diagnostics.

sim2sim

Route safe, source-backed MuJoCo sim2sim validation for the AgiBot X1 DH stand policy, including JIT/model contracts, XML assets, timing, controller mapping, and backend-gated interactive execution.

training

Route X1 DH stand PPO training, configuration inspection, checkpoint lifecycle, and algorithm-only shape checks with explicit Isaac Gym backend limits.

agilerl

Use AgileRL for reinforcement learning workflows: classical RL training, evolutionary HPO, evolvable networks, multi-agent PettingZoo training, offline/bandit data, and LLM fine-tuning.

evolvable-modules

Use AgileRL evolvable modules, networks, architecture configs, custom network wrappers, and mutation-compatible model building blocks.

hpo-and-mutation

Use AgileRL evolutionary HPO, tournament selection, mutation probabilities, mutable hyperparameters, and population evolution safely.

llm-fine-tuning

Use AgileRL LLM fine-tuning and post-training workflows for GRPO, CISPO, GSPO, DPO, SFT, LLM PPO/REINFORCE, vLLM, DeepSpeed, and optional LLM dependencies.

multi-agent-and-wrappers

Use AgileRL multi-agent PettingZoo workflows, MADDPG/MATD3/IPPO, vector envs, wrappers, agent grouping, and multi-agent replay setup.

offline-bandits-data

Use AgileRL offline RL datasets, replay/data conversion, CQL/ILQL, contextual bandits, NeuralUCB/NeuralTS, and BanditEnv workflows.

training-workflows

Use AgileRL classical single-agent training workflows for Gymnasium PPO, DQN, RainbowDQN, DDPG, TD3, replay/rollout buffers, and distributed training setup.

agriculture-knowledge-graph

Operate the Agriculture_KnowledgeGraph agricultural Neo4j graph, Django demo, entity labeling, crawler, and relation-extraction workflows.

crawlers-and-wikidata-pipelines

Operate Agriculture_KnowledgeGraph crawler and Wikidata/weather data-acquisition pipelines safely.

entity-labeling-and-ner

Use THULAC, the agricultural label taxonomy, and the legacy KNN/fastText label workflow for entity recognition and category prediction.

graph-query-and-data-management

Operate the Agriculture KnowledgeGraph Neo4j graph import/query, CSV schemas, hierarchy tree, and vector utilities.

relation-extraction-pipeline

Prepare relation extraction datasets and inspect the TensorFlow PCNN training workflow for Agriculture_KnowledgeGraph.

web-app-service

Operate the Django 1.11 demo application, its routes, forms, and service preflight checks.

ai-data-science-team

Operate the ai-data-science-team package for AI-assisted data loading, EDA, pandas transformations, SQL analysis, H2O/MLflow modeling, multi-agent teams, and Streamlit app workflows.

data-access-and-eda

Use ai-data-science-team data/file loading helpers, direct DataFrame summaries, DataLoaderToolsAgent, EDAToolsAgent, and optional EDA report tools.

dataframe-code-agents

Use ai-data-science-team single-agent pandas code generators for cleaning, wrangling, visualization, and feature engineering workflows.

modeling-and-mlflow

Operate ai-data-science-team H2O AutoML, deterministic model evaluation, MLflow tools, and optional ML dependency checks.

multiagent-and-app-workflows

Compose ai-data-science-team multi-agent workflows and understand the package's Streamlit application patterns.

sql-analysis

Operate ai-data-science-team SQL database querying, metadata inspection, read-only SQL safety, and SQLDatabaseAgent workflows.

ai-optimizer

Route AI-Optimizer reinforcement-learning collection tasks across model-based RL, easy-MARL, offline RL, safe command builders, and repository limitations.

model-based-rl

Guides AI-Optimizer model-based RL baselines, world-model workflows, planning algorithms, and safe MuZero command construction.

multi-agent-rl

Use AI-Optimizer easy-MARL tutorial code for multi-agent RL commands, environments, hyperparameters, and safe MARL extension.

offline-rl

Use AI-Optimizer's offline RL algorithms, d3rlpy-derived APIs, MDPDataset flows, and offline-to-online E2O/PEX workflows safely.

aif360

Use IBM AI Fairness 360 for tabular fairness datasets, metrics, bias mitigation algorithms, sklearn-compatible workflows, subgroup detectors, and explainers.

datasets-and-metrics

Use AIF360 legacy dataset containers and fairness metric classes for tabular protected-group analysis.

detectors-and-explainers

Use AIF360 MDSS and FACTS subgroup detectors plus metric text and JSON explainers.

mitigation-algorithms

Choose and run AIF360 legacy bias mitigation algorithms with correct lifecycle stage, data contracts, optional extras, and metric checks.

sklearn-interface

Use AIF360's preferred sklearn-compatible pandas API for datasets, fairness metrics, scorers, estimators, and pipeline caveats.

aim

Use Aim for experiment tracking SDK instrumentation, local/remote run storage, CLI/UI/server workflows, storage maintenance, and ML framework logging integrations.

cli-and-services

Operate Aim CLI, local UI, remote tracking services, notebook UI, conversion discovery, storage maintenance, and watcher notifications safely.

framework-integrations

Use Aim framework callback integrations, direct tracking fallbacks, and TensorBoard conversion while respecting optional dependency boundaries.