# Alpha-VLLM/LLaMA2-Accessory

An Open-source Toolkit for LLM Development

Repository: https://github.com/Alpha-VLLM/LLaMA2-Accessory
Canonical: https://ross.abutalabs.com/products/llama2-accessory
Homepage: https://llama2-accessory.readthedocs.io/
Language: Python
License: NOASSERTION
License Family: other
Last push: 2025-01-13T06:05:22+00:00

## Health v2 (maintenance only)
Score: 29/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 1, release rhythm 35, longevity 81
- inputs: {"age_days": 1139, "days_push": 597, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2800, forks 176 (observed 2026-08-28T04:07:22.578954+00:00)

## What it is
LLaMA2-Accessory is an open-source Python toolkit for pretraining, finetuning, and deploying large language models and multimodal LLMs, including the SPHINX family of multimodal models. It evolved from LLaMA-Adapter and also includes Large-DiT diffusion model training code.

## Use cases
- finetune a llama-based LLM on custom data
- train a multimodal LLM with vision inputs
- deploy a large language model for inference
- build a visual question answering model
- train a diffusion transformer for text-to-image generation
- evaluate multimodal LLMs on benchmarks

## When to choose
- you want a single toolkit covering LLM pretraining, finetuning, and deployment
- you need to train or finetune multimodal LLMs like SPHINX
- you want to reproduce LLaMA-Adapter or SPHINX research results

## When to avoid
- you need a production-grade serving stack with autoscaling
- you only want to call LLM APIs without training
- you need a lightweight inference-only library

## Facets
- artifact type: framework
- maturity: maintenance
- function: llm-training, llm-inference, machine-learning, deep-learning, rag
- domain: large-language-models, deep-learning, artificial-intelligence, computer-vision
- platform: python
- tags: multimodal-llm, finetuning, sphinx, llama, moe, text-to-image, gpu, linux

## Member repositories
- Alpha-VLLM/LLaMA2-Accessory (main) score 29

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:07:22.578954+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 2026-08-30T08:14:35.197705+00:00, confidence not recorded.
  - readme: https://github.com/Alpha-VLLM/LLaMA2-Accessory (fetched 2026-08-28T04:07:22.578954+00:00, sha 2a757dd56e40)
- Data as of 2026-08-30T08:39:29.467469+00:00.