Skip to content
Robotics

PHASE-TTC-Adapt

TTC по event-камере и RGB

2 min readupdated

Релиз AlvaroSchwiedop для оценки signed time-to-contact (TTC) по causal-наблюдениям event-камеры и RGB-кадрам. Это замороженная оригинальная девятикомпонентная модель PHASE-TTC плюс ансамбль из трёх 688-параметровых quality routers.

Ключевые особенности

  • Считает TTC отдельно по каждой модальности, уточняет оценки во времени и объединяет их в физическом phase-представлении с RGB-коррекцией.
  • Адаптер добавляет causal-расширение bounding box в inverse-TTC координатах; при отсутствии геометрической истории возвращается к native-оценке.
  • Роутеры обучены на 2 608 development-запросах: 2 531 дорожный запрос из 24 записей и 77 запросов Slider-500.
  • Запуск через собственный predict.py; по умолчанию CPU, для CUDA нужны PyTorch 2.10.0 и torchvision 0.25.0.
  • Нужны синхронизированные калиброванные event/RGB-наблюдения и готовые области объектов; автономный детектор — отдельный компонент.

Пример

hf download AlvaroSchwiedop/PHASE-TTC-Adapt --local-dir PHASE-TTC-Adapt
cd PHASE-TTC-Adapt
python -m pip install -r requirements.txt
python predict.py --input examples/synthetic_looming.npz --verify examples/synthetic_expected.json
ПодробнееНа локальной EvTTC-выборке из 434 запросов pooled MAE: 0.82498 → 0.33590 s; relative absolute error: 49.894 → 17.250%.
Это оценка с organizer-style offline negative-output interpolation, а не подтверждённый результат официального лидерборда для адаптации.
Оригинальная PHASE-TTC занимала rank 1 на GarlTTC с 35.80145 Overall MiD; это результат другой, исходной модели.
Paced replay: 29.14 ms p95 на 250 обновлениях — выше бюджета 20 ms; устойчивая работа 50 Hz не показана.
Модель не даёт калиброванных вероятностей столкновения или интервалов неопределённости.
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.

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

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

About the model

AlvaroSchwiedop/PHASE-TTC-Adapt
Author
AlvaroSchwiedop
Category
Robotics
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