Skip to content
Other

Publication Country Classifier

классификатор стран по научным публикациям

2 min readupdated

Модель pub-country-classifier создана на базе архитектуры ModernBERT-base и решает задачу многометочного определения географической принадлежности научных публикаций по их названию и аннотации. На выходе инструмент выдает вероятности для 249 ISO-кодов стран, которые затем фильтруются с помощью откалиброванных порогов. Разработка полезна для исследователей, библиометрических систем и агрегаторов научных статей, которым требуется автоматическая геопривязка материалов.

Ключевые особенности

  • Обучена распознавать упоминания до 249 ISO 3166-1 alpha-2 кодов стран.
  • Использует многометочную сигмоидную «голову» поверх ModernBERT-base.
  • Поддерживает анализ текстов с передачей заголовка на первой строке и аннотации на второй.
  • Поставляется с готовыми файлами порогов (thresholds.json) и кодов для точной фильтрации.

Что можно сделать

  • Автоматически размечать базы данных научных статей и отчетов по странам-участницам исследований.
  • Интегрировать классификатор в пайплайны обработки текстов для фильтрации геотаргетированных публикаций.

Пример

import json, torch
from transformers import AutoTokenizer, AutoModelForSequenceClassification
import huggingface_hub

tok = AutoTokenizer.from_pretrained("Geraldine/pub-country-classifier")
model = AutoModelForSequenceClassification.from_pretrained("Geraldine/pub-country-classifier")
codes = json.load(huggingface_hub.hf_hub_download("Geraldine/pub-country-classifier", "codes.json"))
thr = json.load(huggingface_hub.hf_hub_download("Geraldine/pub-country-classifier", "thresholds.json"))

text = "Water quality governance in rural India\nThis study examines village-level water policy in India ..."
enc = tok(text, truncation=True, max_length=512, return_tensors="pt")
with torch.no_grad():
    probs = torch.sigmoid(model(**enc).logits)[0]
detections = {codes[i]: round(float(probs[i]), 3) for i in range(len(codes)) if float(probs[i]) >= thr[codes[i]]}
print(detections)
Подробнее• micro_f1: 0.9572
• top1_accuracy_on_labelled_rows: 0.868
• precision_at_top1: 0.9927
• Ограничения: корпус на английском языке, тексты длиннее 512 токенов обрезаются, а метки основаны на текстовых упоминаниях
Discuss Telegram 0 46

Will it run?

Runs in F32

Fits without quantisation.

  • F32weights 1.1 GB · ready files1.8 GB
  • BF16weights 0.6 GB1.1 GB
  • FP8 / INT8recommendedweights 0.3 GB0.8 GB
  • 4-bitweights 0.2 GB0.7 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.

    # Geraldine/pub-country-classifier
    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 Geraldine/pub-country-classifier --local-dir ~/models/pub-country-classifier
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to pub-country-classifier
  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 mmBERT-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 1All on Hugging Face

About the model

Geraldine/pub-country-classifier
Author
Geraldine
Category
Other
Released
Posted

On NeyroLearn

Hugging Face

460

Ask AI about this model

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

Comments