Skip to content
Computer vision

amaara-decide-vision

a vision model for fast multiple-choice selection

2 min readupdated

Built by The A-Tech Corporation on top of unsloth/Qwen3.5-0.8B, the compact multimodal model amaara-decide-vision focuses on scoring options rather than generating text or long-form answers. You pass in an image, a question, and a set of answer choices, and the model returns an exact pick with calibrated probabilities for each option in a single pass. It excels at analyzing charts, user interfaces, and maps, returning results in under a second. Developers can use it to quickly automate analytical tasks without the overhead of generating verbose text descriptions.

Key features

  • Selection only: Skips prose generation and directly outputs a probability distribution across the given answer choices.
  • Fast inference: At 0.8B parameters, the model typically processes an image in under a second on a CPU.
  • Built for charts and UI: Trained on chart, diagram, map, and UI element datasets (including rico-widget and mapqa).
  • Unsloth integration: Uses a dedicated QLoRA training setup with custom Clef decision heads.

What you can do

  • Chart analysis: Automatically answer questions about corporate or statistical charts and diagrams.
  • UI inspection: Process mobile and web screenshots to classify UI elements.
  • Map processing: Verify geospatial data and answer yes/no questions about maps.

Example

from unsloth import FastDecisionModel

model, processor = FastDecisionModel.from_pretrained(
    "theatechcorporation/amaara-decide-vision", load_in_4bit=False
)
answers, report = FastDecisionModel.predict(
    model, processor, state="Which region has the highest revenue?",
    questions={"q1": {"type": "choice",
                      "instructions": "Choose the criterion that best answers the question.",
                      "criteria": {"North": "", "South": "", "East": ""}}},
    images=["chart.png"],
)
Read moreDetails
Architecture: based on unsloth/Qwen3.5-0.8B (Apache-2.0), with 38.1M of 699M parameters trained.
Accuracy: achieves 0.848 accuracy on charts (set A) and 0.942 on mapqa maps.
Limitations: geometry (intergps) is not supported; the model is not designed for geometry tasks.
Hardware: tested on an RTX 5060 Ti 16 GB, with support for FP8 quantization and GGUF builds.
Discuss Telegram 0 128

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 1.6 GB · ready files2.3 GB
  • FP8 / INT8recommendedweights 0.8 GB1.4 GB
  • 4-bitweights 0.5 GB1.0 GB

Tip: go 8-bit.FP8 / INT8 is within about a percent of full precision in quality, takes half the memory and runs faster on RTX 40/50 and H100. Full precision only pays off for fine-tuning.

What to get for this model

Massed ComputeOpen Massed Compute

Your own GPU: FP8 / INT8 — from RTX 2060 · BF16 — from RTX 2060

Massed Compute hourly prices, November 2026. The discounted price applies with the coupon at checkout.

Estimate: weights plus working memory. Long videos, high resolutions and long contexts need more.

Run it on Massed Compute

The script downloads the model to the rented machine. How to run it is in the model card.

  1. 1. Rent a machine with a fitting GPU

    Open Massed Compute
  2. 2. Paste the script into the machine's terminal

    Connect over SSH (the Massed Compute panel shows the command) or open a terminal on the machine's desktop.

    HF_TOKEN / CIVITAI_TOKENThe script starts with token lines: paste yours between the quotes. The Hugging Face token is for gated models, the Civitai one for Civitai models.

    # theatechcorporation/amaara-decide-vision
    mkdir -p ~/models
    curl -LsSf https://astral.sh/uv/install.sh | sh
    source "$HOME/.local/bin/env"
    uv tool install "huggingface_hub[cli]"
    hf download theatechcorporation/amaara-decide-vision --local-dir ~/models/amaara-decide-vision
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to amaara-decide-vision
  3. 3. Use it

    The model is in ~/models. How to run it is in the model card (link above).

Stop the machine in the panel when you are done: it bills by the minute while it runs.

How to speed it up

Picked for this model type. Choose your hardware above to see what fits you exactly.

  • FlashAttention 44.0.0b40from November 18, 2026new release

    fastest exact attention on B200

    Kernels for Blackwell (beta). Exact attention without quantisation — fine for everything, LLMs included.

    How to enable: pip install flash-attn-4; vLLM and SGLang pick it up themselves.

    Earlier generations (2)
    • FlashAttention 3 — The Hopper (H100/H200) version with FP8.
    • FlashAttention 2 — The standard for RTX 30/40/50 and A100: exact attention without the memory cost of long contexts.
  • torch.compile2.15.0from October 28, 2026new release

    10–30% for free

    Compiles the model for your GPU. Slower first run, faster afterwards.

    How to enable: model = torch.compile(model); in ComfyUI the TorchCompileModel node.

Model family

  1. Qwen3.5-0.8B-Basefamily root
  2. Qwen3.5-0.8B
  3. Qwen3.5-0.8B
  4. amaara-decide-vision

New posts about versions of Qwen3.5-0.8B-Base (quants, finetunes, LoRAs) will show up under the bell.

Quantised versions 12All on Hugging Face
Fine-tunes 12All on Hugging Face
LoRAs and adapters 9All on Hugging Face
Merges 1All on Hugging Face

About the model

theatechcorporation/amaara-decide-vision
Author
theatechcorporation
Category
Computer vision
Released
Posted

On NeyroLearn

Hugging Face

1280

Ask AI about this model

Answered by Gemini 4 · the chat is kept only in this tab

Comments