Skip to content
Other

QwenRSI Controller v0.1

экспериментальный агент-контроллер

1 min readupdated

QwenRSI Controller v0.1 — экспериментальный компилируемый набор агент-контроллера, который frozen Qwen3.5-4B собрал, модифицируя свой C-каркас. Внутри — реальные разреженные коэффициенты цепей Qwenputer, выбранная RV32-программа, исходники, хеши и измеренные доказательства. Это не обученная языковая модель, не merged-чекпоинт и не LoRA: веса Qwen3.5-4B не входят в репозиторий и загружаются отдельно.

Ключевые особенности

  • В репозитории — разреженные коэффициенты Qwenputer (controller.safetensors), выбранная RV32-программа (controller.elf), исходники и evidence.
  • Загрузка через qwenrsi.circuit.load_bundle, а не AutoModel.from_pretrained; SHA-256 проверяется.
  • Хост по-прежнему отвечает за компиляцию, инициализацию состояния, evaluation, файловый I/O и планирование модели.
  • В измеренном прогоне на RTX 4090 из 3 предложенных версий контроллера принята 1.
  • Это upstream-путь Triton для цепей, а не stock generate() на одной merged-модели.

Пример

from huggingface_hub import snapshot_download
from qwenrsi.circuit import load_bundle

path = snapshot_download("StoneZhang/QwenRSI-Controller-v0.1")
controller = load_bundle(path, device="cuda:0")
print(controller.decide([1, 1, 0, 0, 1, 8, 4]))
ПодробнееБенчмарки: synthetic repair — 0/256 у seed, 216/256 у выбранного, 256/256 у human-written; real Python repair — 4/4 у обоих.
Аудит перезагрузки: 392 токена свободного прогона, 0 расхождений.
Запуск: Linux + GPU-зависимости; измерение на NVIDIA RTX 4090.
Ограничения: улучшение только в симуляции, 4 Python-задачи — авторский пилот, не доказательство широкого улучшения кодинга; автоматического hotspot JIT нет.
Discuss Telegram 0

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 0.0 GB · ready files0.5 GB
  • FP8 / INT8recommendedweights 0.0 GB0.5 GB
  • 4-bitweights 0.0 GB0.5 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.

    # StoneZhang/QwenRSI-Controller-v0.1
    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 StoneZhang/QwenRSI-Controller-v0.1 --local-dir ~/models/QwenRSI-Controller-v0.1
    # The model is in ~/models. How to run it is in the model card (link above).
  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. Qwen3.5-4B-Basefamily root
  2. Qwen3.5-4B
  3. QwenRSI-Controller-v0.1

New posts about versions of Qwen3.5-4B-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 12All on Hugging Face
Merges 1All on Hugging Face

About the model

StoneZhang/QwenRSI-Controller-v0.1
Author
StoneZhang
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

Other

ANT-4B

ANT-4B