# NVlabs/Eagle

Eagle: Frontier Vision-Language Models with Data-Centric Strategies

Repository: https://github.com/NVlabs/Eagle
Canonical: https://ross.abutalabs.com/products/eagle
Homepage: https://nvlabs.github.io/Eagle/
Language: Python
License: Apache-2.0
License Family: permissive
Topics: demo, eagle, gpt4, huggingface, llama, llama3, llava, lmm, lvlm, mllm, llm, large-language-models, nvdia
Last push: 2026-06-24T18:12:11+00:00

## Health v2 (maintenance only)
Score: 64/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 89, release rhythm 35, longevity 56
- inputs: {"age_days": 797, "days_push": 70, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3462, forks 334 (observed 2026-08-28T04:08:05.699800+00:00)

## What it is
Eagle is NVIDIA's family of frontier vision-language models (Eagle, Eagle 2, Eagle 2.5) built with data-centric training strategies, plus LocateAnything, a visual grounding model. The repo provides training, fine-tuning, and inference code along with model weights on Hugging Face.

## Use cases
- run a vision-language model on images and videos
- fine-tune a VLM with LoRA for visual grounding
- understand long videos with a multimodal model
- use a VLM as a backbone for a robotics policy
- build an image and video chat demo
- train a multimodal LLM with data-centric strategies

## When to choose
- you need open weights and code for a state-of-the-art VLM
- you want long-context image/video understanding
- you need a VLM backbone for embodied or robotics research
- you want visual grounding / object localization capabilities

## When to avoid
- you need a lightweight model for CPU-only deployment
- you only want a hosted API without managing GPUs
- you need non-Apache model licensing terms for commercial use

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, llm-training, computer-vision, nlp
- domain: large-language-models, computer-vision, deep-learning, machine-learning, artificial-intelligence
- platform: python
- tags: vision-language-model, multimodal, vlm, video-understanding, visual-grounding, model-weights, nvidia, fine-tuning, lora, huggingface, gpu, linux, docker

## Member repositories
- NVlabs/Eagle (main) score 64

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:05.699800+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:36:53.969295+00:00, confidence not recorded.
  - readme: https://github.com/NVlabs/Eagle (fetched 2026-08-28T04:08:05.699800+00:00, sha a135e13ae145)
  - homepage: https://nvlabs.github.io/Eagle/ (fetched 2026-08-29T09:30:54.659723+00:00, sha 337a2b61db82)
- Data as of 2026-08-30T08:39:29.467469+00:00.