Skip to content
Other

8⃣ HVAC twin v8

LoRA-контроллер для почасовых уставок охлаждения

1 min readupdated

Это LoRA-адаптеры на базе Qwen2.5-3B-Instruct от GeMri98. Модель выбирает почасовые уставки охлаждения по текстовому описанию цифрового двойника офиса 2R2C; качество оценивается симуляцией двойника против точного oracle на grid search.

Ключевые особенности

  • Три адаптера: adapter_v8_sft, adapter_v8_grpo и adapter_v8_sft_grpo.
  • Сравнение SFT, GRPO и связки SFT → GRPO с базовой моделью и rule-based политиками.
  • Метрики: gap_captured, regret, vs_thermostat, cost, viol, valid.
  • Лучший результат у SFT и SFT → GRPO: 75,5% gap_captured, regret 0,849.
  • GRPO от базовой модели слабее: 43,8% gap_captured и regret 1,952.

Пример

Сценарий: на вход подаётся текстовое описание офисного 2R2C-двойника, модель должна выбрать почасовые уставки охлаждения. В таблице карточки вариант SFT then GRPO захватывает 75,5% разрыва до oracle и в 86% случаев лучше термостата 24 °C.

ПодробнееНа 400 эпизодах SFT on oracle и SFT then GRPO дают 75,5% gap_captured, regret 0,849, better_pct 86%, worse_pct 4%, valid 100%.
GRPO from base: gap_captured 43,8%, regret 1,952, better_pct 68%, worse_pct 28%.
База без тюнинга: gap_captured −173,5%, regret 9,490, valid 98%.
Результаты BOPTEST не включены: сервис был недоступен в этом запуске. Оценка только на симуляции двойника, без развёртывания в реальном здании.
Discuss Telegram 0

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.

    # GeMri98/hvac-twin-grpo-v8
    # ===== Tokens: paste yours between the quotes =====
    export HF_TOKEN=""        # huggingface.co/settings/tokens, for gated models
    # ==========================================
    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 GeMri98/hvac-twin-grpo-v8 --local-dir ~/models/hvac-twin-grpo-v8
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to hvac-twin-grpo-v8
  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 hvac-twin-grpo-v8 (quants, finetunes, LoRAs) will show up under the bell.

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

About the model

GeMri98/hvac-twin-grpo-v8
Author
GeMri98
Category
Other
Released
Posted

On NeyroLearn

Hugging Face

00

Ask AI about this model

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

Comments