Skip to content
Other

StereoHand

совместный чекпоинт для стерео-инференса рук

1 min readupdated

Это joint Pi3 inference checkpoint от cskrren для StereoHand: он объединяет temporal MANO на 110 000 шагов, temporal Mask на 40 000 шагов и декодер/голову Pi3 matching pi32. Для совместного инференса используется hand Pi3 register token.

Ключевые особенности

  • Полные восстановленные состояния модели, включая frozen Pi3 и stereo-параметры.
  • Формат stereohand-joint-v1 с отдельными states для mask, mano и matching.
  • Temporal MANO — 110 000 шагов, temporal Mask — 40 000 шагов; одинаковые тензоры делят хранилище, значения и dtypes не меняются.
  • Для запуска нужен код StereoHand и unpack_checkpoint.py; выходная директория должна быть новой.
  • Валидация проверяет сборку чекпоинта, а не новую точность 110k-снапшота.

Пример

import torch
checkpoint = torch.load(
    "stereohand_pi3_mano110k_mask40k_matching_pi32.pt",
    map_location="cpu", mmap=True, weights_only=True,
)
mask_state = checkpoint["states"]["mask"]
mano_state = checkpoint["states"]["mano"]
Подробнее В чекпоинте: temporal MANO — 110 000 шагов, temporal Mask — 40 000 шагов, декодер/голова Pi3 matching pi32.
Формат stereohand-joint-v1; optimizer state отсутствует, значения и dtypes тензоров сохранены.
Валидация проверяет shared Pi3/stereo equality, совместимость matching-backbone и конечность значений.
Это проверка сборки чекпоинта, а не новый бенчмарк точности для 110k-снапшота.
Discuss Telegram 0

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 11 GB · ready files13 GB
  • FP8 / INT8recommendedweights 5.6 GB6.7 GB
  • 4-bitweights 3.2 GB4.0 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 5060 Ti 8GB · BF16 — from RTX 5080

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.

    # cskrren/StereoHand
    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 cskrren/StereoHand --local-dir ~/models/StereoHand
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to StereoHand
  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 StereoHand (quants, finetunes, LoRAs) will show up under the bell.

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

About the model

cskrren/StereoHand
Author
cskrren
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