Skip to content
Other

lead-decision-models-reply-router-merged

роутер ответов лидов без генерации текста

1 min read

Это decision-модель от raoofaltaher — файнтюн LiquidAI/LFM3-1.7B (1.73B) через Unsloth FastDecisionModel и Clef decision head. Она не генерирует текст и не отправляет сообщения, а оценивает заданные варианты и возвращает вероятности: классифицирует ответ на первое sales-письмо по 11 интентам и прогнозирует, нужен ли следующий шаг.

Ключевые особенности

  • 11 интентов ответа лида плюс прогноз необходимости следующего шага.
  • Обучена на 617 синтетических LLM-сгенерированных ответах; тестовая выборка — 124 ответа.
  • Загружается через Unsloth FastDecisionModel; нужен trust_remote_code=True.
  • Пиковое VRAM при обучении — 4.6 GB.
  • Метрики получены на синтетике, а не на реальных CRM-записях и ответах продаж.
  • Лицензия LFM Open License v1.0: коммерческое использование только ниже порога выручки.

Пример

from unsloth import FastDecisionModel

model, tokenizer = FastDecisionModel.from_pretrained(
    "raoofaltaher/lead-decision-models-reply-router-merged",
    trust_remote_code=True,
    load_in_4bit=True,
)
FastDecisionModel.for_inference(model)
ПодробнееТест: accuracy 0.9758, macro-F1 0.9746 против 0.7049 у TF-IDF + logistic regression; ECE 0.0191.
Пиковое VRAM при обучении — 4.6 GB. Два прогона с seed 3407 дали 0.924 и 0.975 macro-F1; тест — 124 ответа.
Решение об исходящем контакте должны принимать policy-код и человек; вероятность — не разрешение.
Лицензия LFM Open License v1.0: коммерческое использование только ниже порога выручки.
Примеры
Discuss Telegram 0 93

Will it run?

Runs in BF16

Fits without quantisation.

  • BF16weights 3.2 GB · ready files4.0 GB
  • FP8 / INT8recommendedweights 1.6 GB2.3 GB
  • 4-bitweights 0.9 GB1.5 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.

    # raoofaltaher/lead-decision-models-reply-router-merged
    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 raoofaltaher/lead-decision-models-reply-router-merged --local-dir ~/models/lead-decision-models-reply-router-merged
    # 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. LFM3-1.7B-Basefamily root
  2. LFM3-1.7B
  3. lead-decision-models-reply-router-merged

New posts about versions of LFM3-1.7B-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 9All on Hugging Face

About the model

raoofaltaher/lead-decision-models-reply-router-merged
Author
raoofaltaher
Category
Other
Released
Posted

On NeyroLearn

Hugging Face

930

Ask AI about this model

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

Comments