Skip to content
Language models

Gomoku

Dense Llama — крошечная LLM для гомоку 15×15

2 min readupdated

Это специализированная модель для гомоку на базе LlamaForCausalLM. Она получает полное состояние доски и генерирует координату хода; всего 10 809 984 параметра в F32 без квантизации. Поддерживает Transformers и нативный Ollama, но не является универсальным чат-ботом или шахматной моделью.

Ключевые особенности

  • 3 слоя, hidden 128, 4 головы внимания, FFN 2048, контекст 2048 токенов.
  • В комплекте веса safetensors, GGUF F32, Modelfile для Ollama и терминальный клиент.
  • На тесте Dense даёт 153/224 целевых попаданий и 222/224 легальных хода.
  • Dense против MoE: 9–7 в прямых партиях; медиана в Ollama — 174 мс против 104 мс.
  • Модель не вызывает движок, не ищет и не фильтрует нелегальные ходы — ошибки возможны.

Пример

from transformers import AutoModelForCausalLM, AutoTokenizer
from pointer_protocol import messages

tok = AutoTokenizer.from_pretrained('.', local_files_only=True)
model = AutoModelForCausalLM.from_pretrained('.', local_files_only=True).eval()
prompt = tok.apply_chat_template(messages([0] * 225), tokenize=False, add_generation_prompt=True)
inputs = tok(prompt, add_special_tokens=False, return_tensors='pt')
out = model.generate(**inputs, max_new_tokens=4, do_sample=False)
print(tok.decode(out[0, inputs['input_ids'].shape[1]:], skip_special_tokens=True))
ПодробнееТест: 153/224 попаданий в цель и 222/224 легальных хода у Dense; у MoE — 151/224 и 222/224.
Прямые партии Dense vs MoE: 9–7. Медиана в Ollama на CPU, 2 потока, F32: 174 мс и 104 мс.
Это экспериментальная модель: не гарантирует легальный ход и не заменяет поиск или движок.
Обучение — смесь reward и supervised, 256 шагов; лицензия Apache-2.0.
Discuss Telegram 0 992

Will it run?

Runs in GGUF F32

The most precise of the ready quantised files fits.

  • GGUF F32weights 0.1 GB · ready files0.7 GB
  • F32weights 0.0 GB · ready files0.6 GB
  • BF16weights 0.0 GB0.6 GB
  • FP8 / INT8recommendedweights 0.0 GB0.6 GB
  • 4-bitweights 0.0 GB0.6 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 · GGUF F32 — from RTX 2060

Massed Compute hourly prices, November 2026. The discounted price applies with the coupon at checkout.

Language models: with a 8192-token context.

Estimate: weights plus working memory. Long videos, high resolutions and long contexts need more.

Run it on Massed Compute

A ready script for the machine we advise above: paste it and the model downloads and starts.

  1. 1. Rent A30

    $0.26 an hour with the coupon. Pick the Ubuntu image when you create the machine.

    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.

    # l898974766/gomoku — A30, FP8 / INT8
    # NVIDIA driver: CUDA 13 needs 580+. If older or missing, the current nvidia-open is installed (keep yours: KEEP_DRIVER=1)
    DRV=$(nvidia-smi --query-gpu=driver_version --format=csv,noheader 2>/dev/null | head -1 | cut -d. -f1)
    case "$DRV" in ""|*[!0-9]*) DRV=0 ;; esac
    if [ "$DRV" -lt 580 ] && [ -z "${KEEP_DRIVER:-}" ]; then
      . /etc/os-release && \
      curl -fsSLO "https://developer.download.nvidia.com/compute/cuda/repos/ubuntu${VERSION_ID//./}/x86_64/cuda-keyring_1.1-1_all.deb" && \
      sudo dpkg -i cuda-keyring_1.1-1_all.deb && rm -f cuda-keyring_1.1-1_all.deb && \
      sudo apt-get update && \
      sudo DEBIAN_FRONTEND=noninteractive apt-get install -y nvidia-open || \
      { echo "Could not install the NVIDIA driver: see the error above. To keep the current one, run with KEEP_DRIVER=1."; exit 1; }
      echo "NVIDIA driver updated. Reboot the machine (sudo reboot) and run this same script again."
      exit 0
    fi
    curl -LsSf https://astral.sh/uv/install.sh | sh
    source "$HOME/.local/bin/env"
    uv venv ~/vllm --python 3.12 && source ~/vllm/bin/activate
    uv pip install vllm
    vllm serve l898974766/gomoku --max-model-len 8192 --quantization fp8
    Add more LoRAs to gomoku
  3. 3. Use it

    The server listens on port 8000 and speaks the OpenAI API: http://localhost:8000/v1 on the machine. From your computer: ssh -L 8000:localhost:8000 user@machine-IP.

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.

  • MTP (multi-token prediction)0.34.0from November 16, 2026new release

    1.5–2× generation speed

    If the model has MTP layers (DeepSeek-V3/R1, Qwen3-Next, GLM, MiMo and others), it predicts several tokens per step itself.

    How to enable: vLLM / SGLang: --speculative-config '{"method": "mtp", "num_speculative_tokens": 1}'. Check the model card.

    Earlier generations (2)
    • EAGLE-3 — Speculative decoding with a small draft head. For models without their own MTP layers.
    • EAGLE-2 / Medusa — Earlier draft-head generations. Still work if that is all you find.
  • FP8 KV cache0.34.0from November 16, 2026new release

    2× more context in the same memory

    Keeps the attention cache in 8 bits, so long contexts fit where they did not.

    How to enable: vLLM: --kv-cache-dtype fp8; llama.cpp: -ctk q8_0 -ctv q8_0.

  • vLLM / SGLang0.34.0from November 16, 2026new release

    many times more tokens/s on a server

    Server engines: batching, FlashAttention, FP8 and speculative decoding out of the box.

    How to enable: vllm serve <model> — an OpenAI-compatible API.

    Earlier generations (1)
    • llama.cpp — GGUF models on any GPU, Mac or even CPU; some layers can stay in RAM.

Model family

New posts about versions of gomoku (quants, finetunes, LoRAs) will show up under the bell.

Quantised versions 1All on Hugging Face

About the model

l898974766/gomoku
Author
l898974766
Category
Language models
Released
Posted

On NeyroLearn

Hugging Face

9920

Ask AI about this model

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

Comments

Language models

FLY-SFLY

FLY-SFLY