AA

aac pictogram retriever ptbr

AacGGUF by PapuguinhoUpdated 2026-10-20
Download on Hugging Face
Parameters155M
Context2K
Layers24
Architecturegemma-embedding
Smallest quant0.1 GB
Quick answer

aac pictogram retriever ptbr needs about 0.8 GB of memory at Q4_K_M (a 0.1 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 4 quants, from 0.1 GB (Q4_K_M) to 0.3 GB (F16).

AAC Pictogram Retriever (pt-BR) – compact embedding model for ARASAAC pictogram search

AAC Pictogram Retriever is a 0.155B parameter model built for Brazilian‑Portuguese pictogram retrieval in AAC applications. It supports retrieval and reranking of ~14k ARASAAC pictograms. The model runs comfortably on CPU and fits in under 200 MB of RAM at Q8_0.

Good at

  • Pictogram retrieval in AAC
  • Short‑query semantic search
  • Brazilian‑Portuguese embeddings

Limits

  • Not for general‑purpose embeddings
  • No support for languages other than Brazilian Portuguese
  • No vision or multimodal capabilities

Best for

  • Building communication boards
  • Therapist‑child interaction tools
  • Pictogram search in AAC apps

Who it’s for. Developers and therapists creating AAC tools that require fast, local embedding of Brazilian‑Portuguese queries. The model is ideal for low‑resource environments and self‑hosted deployments.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q4_K_M0.1 GB 0.7 GB
Fits Get file →
Q6_K0.1 GB 0.7 GB
Fits Get file →
Q8_00.2 GB 0.7 GB
Fits Get file →
F160.3 GB 0.9 GB
Fits Get file →

For 8 GB VRAM, Q4_K_M or Q6_K are suitable, each using about 0.8 GB of memory at 8k context. Q8_0 uses 0.9 GB and F16 uses 1 GB. Choose the smallest quant that meets accuracy needs.

What GPU do you need for aac pictogram retriever ptbr?

These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.

  • 4 GB VRAM: the best quant that fits is F16 (0.3 GB file, about 1.0 GB in use).

Which aac pictogram retriever ptbr quant should you choose?

Q4_K_M is the usual starting point for aac pictogram retriever ptbr: it balances file size and quality well. Here it is a 0.1 GB file.

The Q8_0 file is 0.2 GB, about 1.5 times the size of Q4_K_M. It keeps the weights closer to the original, which only matters if you have the memory to spare.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturegemma-embedding
Parameters155M
Layers24
Trained context2,048 tokens
KV cache (fp16)24 KB per token

Source

Quantized byPapuguinho
FamilyAac
Licensegemma
Downloads (HF)2,366
Last updated2026-10-20

Frequently asked questions about aac pictogram retriever ptbr

How much VRAM does aac pictogram retriever ptbr need?

About 0.8 GB at Q4_K_M with an 8K context window. That is the 0.1 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.

Can I run aac pictogram retriever ptbr on an 8 GB GPU?

Yes, with F16: the 0.3 GB file needs about 1.0 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of aac pictogram retriever ptbr GGUF?

It depends on the quant: Q4_K_M is 0.1 GB, Q6_K is 0.1 GB, Q8_0 is 0.2 GB, F16 is 0.3 GB.

What context length does aac pictogram retriever ptbr support?

The GGUF header lists a trained context length of 2,048 tokens (2K). The KV cache grows with the context you use: about 0.2 GB at 8K and 0.1 GB at 2K.

Where can I download aac pictogram retriever ptbr GGUF?

The files are on Hugging Face in the Papuguinho/aac-pictogram-retriever-ptbr-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

What license does aac pictogram retriever ptbr use?

Hugging Face lists the license as gemma. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.

Can I run AAC Pictogram Retriever on a CPU only machine?

Yes, the model is designed for CPU inference and does not require a GPU.

What is the maximum context length supported?

The model was trained with 2048 tokens and supports up to that context length.

Is the model multilingual?

No, it is specifically tuned for Brazilian‑Portuguese and not intended for other languages.

Similar models

Browse all →

More from Papuguinho

All by Papuguinho →

More Aac models

All Aac →

Other tiny models

All tiny models →