# amalia-llm/amalia architecture

> Architecture diagram of the GitHub repository amalia-llm/amalia, generated by GitDiagram from its file tree, README and a sample of source files (last updated 2026-11-21). It is an AI-made overview: check details against the source.

- Interactive diagram: https://gitdiagram.com/amalia-llm/amalia
- Repository: https://github.com/amalia-llm/amalia

## Overview

This repository is an AMALIA project landing page, not the implementation of its models or pipelines. It describes an open European Portuguese language-model ecosystem and documents the main user workflow: deploy an AMALIA model with vLLM, send a chat-completion request, and receive an answer. The graph also preserves the documented data, training, post-training, vision-language, and evaluation capabilities. No source excerpts are available, so subsystem paths and relationships rely on README.md; the README does not establish detailed interactions among the research components.

Main entry points: [`README.md`](https://github.com/amalia-llm/amalia/blob/main/README.md) (AMALIA models)

## Diagram (Mermaid)

```mermaid
flowchart TD

subgraph group_serving["Model serving"]
  node_vllm["vLLM serving"]
  node_amalia_models[("AMALIA models<br/>[README.md]")]
end

subgraph group_data["Data processing"]
  node_data_pipelines["Data pipelines<br/>[README.md]"]
end

subgraph group_training["Model development"]
  node_pretraining["Pretraining<br/>[README.md]"]
  node_posttraining["Language post-training<br/>[README.md]"]
  node_vl_finetuning["Vision-language tuning<br/>[README.md]"]
end

subgraph group_evaluation["Evaluation"]
  node_lm_evaluation["Language evaluation<br/>[README.md]"]
  node_vl_evaluation["Vision-language evaluation<br/>[README.md]"]
  node_benchmarks["Portuguese benchmarks<br/>[README.md]"]
end

node_user(("Model user"))

node_user -->|"sends request"| node_vllm
node_vllm -->|"serves model"| node_amalia_models
node_vllm -->|"returns answer"| node_user

click node_amalia_models "https://github.com/amalia-llm/amalia/blob/main/README.md"
click node_data_pipelines "https://github.com/amalia-llm/amalia/blob/main/README.md"
click node_pretraining "https://github.com/amalia-llm/amalia/blob/main/README.md"
click node_posttraining "https://github.com/amalia-llm/amalia/blob/main/README.md"
click node_vl_finetuning "https://github.com/amalia-llm/amalia/blob/main/README.md"
click node_lm_evaluation "https://github.com/amalia-llm/amalia/blob/main/README.md"
click node_vl_evaluation "https://github.com/amalia-llm/amalia/blob/main/README.md"
click node_benchmarks "https://github.com/amalia-llm/amalia/blob/main/README.md"

classDef toneNeutral fill:#f8fafc,stroke:#334155,stroke-width:1.5px,color:#0f172a
classDef toneBlue fill:#dbeafe,stroke:#2563eb,stroke-width:1.5px,color:#172554
classDef toneAmber fill:#fef3c7,stroke:#d97706,stroke-width:1.5px,color:#78350f
classDef toneMint fill:#dcfce7,stroke:#16a34a,stroke-width:1.5px,color:#14532d
classDef toneRose fill:#ffe4e6,stroke:#e11d48,stroke-width:1.5px,color:#881337
classDef toneIndigo fill:#e0e7ff,stroke:#4f46e5,stroke-width:1.5px,color:#312e81
classDef toneTeal fill:#ccfbf1,stroke:#0f766e,stroke-width:1.5px,color:#134e4a
class node_vllm,node_amalia_models,node_user toneBlue
class node_data_pipelines toneAmber
class node_pretraining,node_posttraining,node_vl_finetuning toneMint
class node_lm_evaluation,node_vl_evaluation,node_benchmarks toneRose
```

## Components

### Model serving

- **vLLM serving**
- [AMALIA models](https://github.com/amalia-llm/amalia/blob/main/README.md): `README.md`

### Data processing

- [Data pipelines](https://github.com/amalia-llm/amalia/blob/main/README.md): `README.md`

### Model development

- [Pretraining](https://github.com/amalia-llm/amalia/blob/main/README.md): `README.md`
- [Language post-training](https://github.com/amalia-llm/amalia/blob/main/README.md): `README.md`
- [Vision-language tuning](https://github.com/amalia-llm/amalia/blob/main/README.md): `README.md`

### Evaluation

- [Language evaluation](https://github.com/amalia-llm/amalia/blob/main/README.md): `README.md`
- [Vision-language evaluation](https://github.com/amalia-llm/amalia/blob/main/README.md): `README.md`
- [Portuguese benchmarks](https://github.com/amalia-llm/amalia/blob/main/README.md): `README.md`

### Other components

- **Model user**

## Connections

- Model user → vLLM serving: sends request (evidence: `README.md`)
- vLLM serving → AMALIA models: serves model (evidence: `README.md`)
- vLLM serving → Model user: returns answer (evidence: `README.md`)
