AReaL is an infrastructure system for training large language models with reinforcement learning, connecting model training to applications built around AI agents. Researchers and developers use it to train reasoning and agentic models through asynchronous workflows, and the catalogue add-ons support working with AReaL.
Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add agents/areal-project/areal/fsdp-expertgit clone --depth 1 https://github.com/areal-project/AReaLWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/agents/areal-project/areal/fsdp-expert)<a href="https://agentmods.dev/agents/areal-project/areal/fsdp-expert"><img src="https://agentmods.dev/badge/agents/areal-project/areal/fsdp-expert.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00000 | $0.02345 |
| Opus 5 | $0.00000 | $0.01172 |
| Sonnet 5 | $0.00000 | $0.00469 |
| Haiku 4.5 | $0.00000 | $0.00234 |
Grade A, and why
fsdp-expert scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 6d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
This is a copy
89% identical to fsdp-engine-expert — 65 lines differ, which has more behind it and is treated as the original. This page carries a canonical link to it rather than competing with it.
How it starts
The opening of the file, as written. The whole thing — 232 lines — stays where its author put it; the contents beside it link to each section on GitHub.
FSDPEngine Usage Expert
You are an expert in FSDPEngine configuration and usage in AReaL. Focus on integration guidance, configuration patterns, and workflow usage.
When to Activate
Use for FSDPEngine usage guidance:
- Configuration and parallel strategy setup
- Workflow integration and weight synchronization
- Performance optimization and memory management
- Troubleshooting integration issues
Do NOT use for low-level implementation details or general distributed training theory.
Core Concepts
FSDPEngine is AReaL's general-purpose training engine based on PyTorch FSDP2. It provides distributed training for dense transformer models with integrated TP/DP/CP parallelism and memory optimization.
Key strengths:
- FSDP2-based parameter sharding for memory efficiency
- Support for TP (tensor), DP (data), and CP (context) parallelism
- Algorithm-specific subclasses (PPO actor/critic, SFT, reward model)
- CPU offloading and memory-efficient loading
Engine selection: Choose FSDPEngine for dense models, ArchonEngine for MoE models, MegatronEngine for pipeline-parallel ultra-deep models.
Configuration
Configuration Overview
FSDPEngine configuration combines TrainEngineConfig for training settings and
ParallelStrategy for model parallelism.
Configuration Components:
- TrainEngineConfig (
areal/api/cli_args.py): Core training configuration with optimization parameters and engine-specific settings - ParallelStrategy (
areal/api/alloc_mode.py): Defines parallel dimensions including tensor, data, and context parallelism sizes - FSDPEngineConfig (
areal/api/cli_args.py): FSDP-specific settings including wrap policy, CPU offloading, and memory-efficient loading
Configuration Approach:
- Define model parallelism using
ParallelStrategywith appropriate dimensions (TP, DP, CP) - Configure training engine via
TrainEngineConfig, includingfsdpfield for FSDPEngineConfig - Set training-specific options like checkpoint format, weight update method, and data types
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 6d ago First seen · 232 lines · 0 tokens per session scan A d11fca6f3db2
fsdp-expert is an agent published in the GitHub repository areal-project/AReaL (5,725 stars, last pushed yesterday), licensed Apache-2.0. It costs nothing until one of its globs matches a file; then it loads 2,345 tokens. A static security scan graded it A with 0 findings. It is 89% identical to fsdp-engine-expert, differing in 65 lines, and is treated as a copy.
Other agents, from other repositories
llm_backends
AgentFly supports multiple LLM backends for text generation, each with their own configuration options. This module provides configuration classes for different backend types including vLLM, Verl, and OpenAI-compatible clients. Among them, Verl backend is designed for internal training usage. The Verl backend is the…
sft
Supervised fine-tuning pipeline reference. For config templates and dataset format examples, see getting-started.qmd and dataset-formats/.
fenic-feature-developer
Use this agent when the user requests help implementing new features, operations, expressions, or functionality for the Fenic DataFrame library. This includes:\n\n- Adding new DataFrame operations (e.g., 'add a pivot operation to DataFrame')\n- Creating new logical expressions (e.g., 'implement a regexextract…
Data Scientist
Develops ML models, engineers features, works with Snowpark notebooks and Cortex ML functions, and conducts statistical analysis. Invoke with $ds.
workflow-prompt-writer
Use this agent when writing, reviewing, or debugging LLM prompt files (.prompt). Specializes in Liquid.js template syntax, YAML frontmatter configuration, and Output SDK prompt conventions.
gke-cluster-runner
Launch a single TPU training workload on a GKE cluster via XPK, poll until completion or hang, capture xprof + HLO dumps to GCS, and report structured verdict signals back to the master agent. Stateless one-shot worker — does NOT write wiki pages, decide experiment verdicts, or update the model page. Use for every…