eagle3-new-model

A configuration workflow for adding a new model checkpoint to the offline EAGLE3 pipeline. EAGLE3 is a pipeline that prepares data, extracts model hidden states, trains, and benchmarks the model.

In plain words
What is it for?
Use it when a model has no YAML launcher file yet and you need to run it through EAGLE3. It covers text, multimodal, custom-code, and sliding-window model cases according to the documented backend choices.
Why use it?
It helps choose an existing configuration to adapt, the suitable hidden-state extraction backend, and the required GPU settings. This avoids building the four-stage launcher configuration from scratch.

Skill for Claude CodeCodex

Part of the modelopt plugin — 18 skills shipped together

Install

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.

agentmods
npx agentmods add skills/nvidia/model-optimizer/eagle3-new-model
Any agent
npx skills add NVIDIA/Model-Optimizer --skill eagle3-new-model
Clone the repo
git clone --depth 1 https://github.com/NVIDIA/Model-Optimizer

Made for: Claude Code, Codex.

Or install modelopt, the plugin that ships this one along with the rest of its 18 skills.

Per session 98 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 665 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00098 $0.00665
Opus 5 $0.00049 $0.00332
Sonnet 5 $0.00020 $0.00133
Haiku 4.5 $0.00010 $0.00067

Measured 3d ago against content hash d47d39626420, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

eagle3-new-model 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 3d 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.

plugins/modelopt/skills/eagle3-new-model/SKILL.md · 47 lines

How it starts

The opening of the file, as written. The whole thing — 47 lines — stays where its author put it; the contents beside it link to each section on GitHub.

EAGLE3 New Model Configuration

Create tools/launcher/examples/<Org>/<Model>/hf_offline_eagle3.yaml by copying the closest existing example and adapting it. Pick a reference with the same shape as the target (dense vs MoE, similar size) from tools/launcher/examples/ — e.g. the Qwen3-8B config for a dense model.

The pipeline is a 4-task config (task_0 data synthesis → task_1 hidden-state dump → task_2 train → task_3 benchmark). The task structure, args, containers, and GPU/node sizing are all visible in the existing examples — infer them from a reference rather than hand-rolling. This file documents only the two things that are not obvious from the examples: which dump backend to pick, and the model-specific gotchas.

Choosing the task_1 hidden-state dump backend

Backend Script When to use
vLLM common/eagle3/dump_offline_data_vllm.sh Default. Broad coverage via vLLM's native hidden-state extractor.
HF common/eagle3/dump_offline_data_hf.sh VLMs / multimodal, custom-code models, sliding-window attention (TRT-LLM can't serve these).
TRT-LLM common/eagle3/dump_offline_data.sh Pure-text models with TRT-LLM support; pass --tp <TP> and --moe-ep <EP>.

Rule of thumb: HF if the model is a VLM or uses sliding-window attention; vLLM otherwise. TRT-LLM only when you specifically want its kernels for a supported plain-text model.

Model-specific adjustments

These are the non-obvious knobs that vary per model:

Situation What to change
Requires --trust-remote-code Add to task_0 vLLM args (before the -- separator) and to task_3 benchmark args
MoE with large expert hidden dim Increase intermediate_size in eagle_config.json to match moe_intermediate_size
Custom tokenizer (e.g. tiktoken) Set TIKTOKEN_RS_CACHE_DIR env var in task_0 and task_1

After adapting the config, preview it with --dryrun before submitting.

Read the full file on GitHub · 47 lines

Changes

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.

  1. 3d ago First seen · 47 lines · 98 tokens per session scan A d47d39626420

Subscribe to this mod's changes

eagle3-new-model is a skill published in the GitHub repository NVIDIA/Model-Optimizer (3,675 stars, last pushed yesterday), licensed Apache-2.0. It adds 98 tokens to every session and 665 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other skills, from other repositories

agent-platform-model-registry

Agent Platform Model Registry Management. Use when you need to upload, list, describe, update, or delete machine learning models (and their versions) in the Agent Platform Model Registry. Don't use for model training, model deployment to endpoints, or managing non-Agent Platform models.

google/skills · 60 tokens

bigquery-ai-ml

Leverages BigQuery's built-in machine learning and GenAI capabilities for advanced data analytics. Use when you need to write SQL queries that perform time-series forecasting, predict values, detect outliers or anomalies, find key drivers, perform semantic search or vector search, classify text, calculate similarity…

google/skills · 104 tokens

ondb

A logical analysis and reasoning tool for AI. Use when decomposing documents into structured knowledge, querying entities and relations, validating consistency, or indexing files. Trigger on "remember", "what do I know about", "link X to Y", "show dependencies", "analyze this document", entity CRUD, or cross-skill…

x-cmd/x-cmd · 71 tokens

neuron-evaluation-engineer

Create and run AI evaluations with datasets, assertions, and output drivers in Neuron AI. Use this skill whenever the user mentions evaluation, testing AI systems, creating evaluators, dataset-driven testing, assertion-based validation, or wants to measure AI system performance. Also trigger for tasks involving…

neuron-core/neuron-ai · 77 tokens

mixed-precision

Use FP16/BF16 mixed precision to accelerate training and reduce memory. Use when optimizing GPU performance.

aiming-lab/AutoResearchClaw · 25 tokens

sparse-autoencoder-training

Provides guidance for training and analyzing Sparse Autoencoders (SAEs) using SAELens to decompose neural network activations into interpretable features. Use when discovering interpretable features, analyzing superposition, or studying monosemantic representations in language models.

Orchestra-Research/AI-Research-SKILLs · 58 tokens