Skip to content
Other

GeoTrace photo year

оценка года снимка по стилю

2 min readupdated

Это классификатор изображений от davidlnk3, который оценивает год фотографии по её фотографическому стилю. Модель используется в инструменте time_from_objects проекта GeoTrace и воспроизводит подход статьи «Blind Dates». Архитектура — логистическая регрессия на 50 классов (1950–1999) поверх замороженных эмбеддингов OpenCLIP.

Ключевые особенности

  • Backbone ViT-B-32-quickgelu / laion400m_e32 не дообучается: используется сырой encode_image без L2-нормализации, 512-d.
  • Голова — LogisticRegression(random_state=0, max_iter=5000) из scikit-learn 1.9.1, по одному классу на год.
  • Веса головы опубликованы как массивы в safetensors: weight 50×512 и bias; для инференса не нужен scikit-learn.
  • Обучение на De Boer Scene Detection: 31 892 train / 7 974 test, плюс 500 фото DEW; split взят у авторов статьи.
  • Модель читает стиль снимка — зерно, цветопередачу, процесс, кадрирование, — а не содержание сцены.
  • Классы ограничены 1950–1999: на фото вне диапазона модель всё равно вернёт год из него.

Пример

import numpy as np
from safetensors.numpy import load_file

t = load_file('model.safetensors')
logits = embedding @ t['weight'].T + t['bias']
year = int(t['classes'][logits.argmax()])
# embedding — сырой encode_image, 512-d
ПодробнееНа тесте De Boer (7 974 изображения): accuracy 0.1366, MAE 6.71 года; без DEW — 6.65.
На 400 отложенных DEW MAE 13.29 против 13.55 без добавки.
Вне диапазона 1950–1999 вывод модели недействителен как дата снимка.
Обучено 2026-11-10 на Python 3.12.15; не предназначено для идентификации людей.
Discuss Telegram 0 15

Will it run?

Runs in F32

Fits without quantisation.

  • F32weights 0.0 GB · ready files0.5 GB
  • BF16weights 0.0 GB0.5 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 · F32 — 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.

    # davidlnk3/geotrace-photo-year
    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 davidlnk3/geotrace-photo-year --local-dir ~/models/geotrace-photo-year
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to geotrace-photo-year
  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 geotrace-photo-year (quants, finetunes, LoRAs) will show up under the bell.

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

About the model

davidlnk3/geotrace-photo-year
Author
davidlnk3
Category
Other
Released
Posted

On NeyroLearn

Hugging Face

150

Ask AI about this model

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

Comments

Other

AI Photo

AI Photo