Skip to content
Computer vision

LocHeal Verdict 2B

VLM для самовосстановления веб-локаторов

1 min readupdated

LocHeal Verdict 2B — специализированная System 1 Typed-Decision vision-language model на базе LiquidAI/d1-2B (LFM3-VL-2B tri-hybrid), около 1.92B параметров. Она предназначена для automated self-healing локаторов и multimodal DOM grounding в Playwright/Selenium: получает SoM-скриншот и телеметрию сломанного локатора, затем за один forward pass оценивает живые DOM-кандидаты.

Ключевые особенности

  • Candidate Selection: модель не генерирует XPath/CSS, а выбирает индекс живого DOM-кандидата через SlotReader.
  • Visual + Semantic Fusion: учитывает SoM-скриншот, позицию и смысл элемента, даже если HTML сильно изменился.
  • Явный отказ Unknown: обучена на удалении элементов и блокировках, чтобы не кликать ложную кнопку.
  • Работает с description и без: curriculum 50/50, без описания опирается на broken_locator, variable_name и intended_action.
  • Merged LoRA r=16, alpha=32; temperature scaling в calibration.json; веса bfloat16 плюс бэкап адаптера.

Пример

engine = slots_mod.SlotReader(model, processor)
result = engine.decide(
    state=broken_locator_state,
    questions={"healed_candidate": question},
    images=[som_screenshot]
)
print(result)
ПодробнееБаза: LiquidAI/d1-2B, 1.92B параметров; vision tower 164M, бюджет 896×896.
В продакшене при Unknown или confidence < 0.75 тест падает, иначе локатор синтезируется из выбранного DOM-узла.
Лицензия наследуется от LFM Open License v1.0.
Бенчмарков в карточке нет; калибровка — LBFGS temperature scaling по бакетам low/mid/high.
Discuss Telegram 1 61

Will it run?

Working out the memory needs

The bot is still collecting this model's weight sizes; the estimate will appear here in a few minutes. Meanwhile you can rent a GPU with headroom.

Massed ComputeOpen Massed Compute

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.

    # kinasr/locheal-verdict-2b
    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 kinasr/locheal-verdict-2b --local-dir ~/models/locheal-verdict-2b
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to locheal-verdict-2b
  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

New posts about versions of locheal-verdict-2b (quants, finetunes, LoRAs) will show up under the bell.

No known relatives of this model yet. All derivatives on Hugging Face

About the model

kinasr/locheal-verdict-2b
Author
kinasr
Category
Computer vision
Released
Posted

On NeyroLearn

Hugging Face

611

Ask AI about this model

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

Comments

Computer vision

JEV-27B-VL

JEV-27B-VL
Computer vision

UAV-PUR-8B

UAV-PUR-8B