Skip to content
Other

LFM2.5-24B-A2B · Harbor · pi

LoRA-адаптер для агента pi на анализе данных

1 min read

Это LoRA-адаптер для LiquidAI/LFM2.5-24B-A2B от harness-hub. Обучен с подкреплением через AsyncGRPO в TRL как модель для кодинг-агента pi на задачах анализа данных: каждый rollout — полная сессия pi в удалённом sandbox с shell и Python.

Ключевые особенности

  • LoRA r=32, alpha=64, все linear-слои; обучение — AsyncGRPO (TRL, experimental).
  • Решает задачи примерно так же часто, как базовая модель, но тратит на 39% меньше tool calls на задачах, решённых обеими.
  • Обучен и оценён внутри harness pi; для vLLM нужен --tool-call-parser lfm2.
  • Подходит, если нужен более дешёвый и быстрый агент при сопоставимой точности.
  • Есть GGUF-версия адаптера для llama.cpp, но она не переоценивалась.

Пример

vllm serve LiquidAI/LFM2.5-24B-A2B \
    --enable-lora --max-lora-rank 32 \
    --lora-modules lfm-harbor-pi=harness-hub/LFM2.5-24B-A2B-harbor-pi-lora \
    --enable-auto-tool-choice --tool-call-parser lfm2 --reasoning-parser qwen3 \
    --max-model-len 131072
ПодробнееПодробнее
Оценка: pass@1 −4.2 пункта — в пределах шума ±5 на 100 задачах; tool calls −39% на задачах, решённых обеими моделями.
Обучение: 100 шагов, lr 1e-5, 1× NVIDIA B300, первые 1000 задач SmolDataEnvs-harbor-train.
Ограничения: маленькая оценка, только семейство задач анализа данных SmolDataEnvs и только harness pi; с другим агентом поведение может быть как у базовой модели или хуже.
GGUF-версия не переоценивалась.
Примеры
Discuss Telegram 10 22

Will it run?

Working out the memory needs

The bot is still collecting this model's weight sizes; the estimate will appear here in a few minutes. Meanwhile you can rent a GPU with headroom.

Massed ComputeOpen Massed Compute

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.

    # harness-hub/LFM2.5-24B-A2B-pi
    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 harness-hub/LFM2.5-24B-A2B-pi --local-dir ~/models/LFM2.5-24B-A2B-pi
    # The model is in ~/models. How to run it is in the model card (link above).
    Add more LoRAs to LFM2.5-24B-A2B-pi
  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 LFM2.5-24B-A2B-pi (quants, finetunes, LoRAs) will show up under the bell.

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

About the model

harness-hub/LFM2.5-24B-A2B-pi
Author
harness-hub
Category
Other
Released
Posted

On NeyroLearn

Hugging Face

2210

Ask AI about this model

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

Comments