Unsloth 发布 Qwen3.8 系列动态量化 v3.1 更新

原标题：Unsloth Qwen3.8-27B Quantizations – Dynamic v3.1 Update

58AI 研判分

Unsloth 为 Qwen3.8-27B 及 2.4T-A95B 等模型发布 Dynamic v3.1 量化版本，提供 UD-Q4_K_XL、IQ1_S 等多档 GGUF 权重，并配套一键安装脚本与 llama.cpp 本地部署命令。该更新降低了超大 MoE 模型在消费级与低比特环境下的本地运行门槛。

关联知识实体

术语动态量化术语GGUF 工具llama.cpp 模型Qwen3.8-2.4T-A95B 公司Unsloth 模型Qwen3.8-27B

报道正文

Was this helpful?

BASH

curl -fsSL https://unsloth.ai/install.sh | sh

文本

irm https://unsloth.ai/install.ps1 | iex

文本

brew install --cask unsloth

文本

brew upgrade --cask unsloth

文本

winget install --id UnslothAI.Unsloth -e

文本

winget upgrade --id UnslothAI.Unsloth -e

文本

unsloth run --model unsloth/Qwen3.8-27B-GGUF:UD-Q4_K_XL
    --temp 1.0 \
    --top-p 0.95 \
    --top-k 20 \
    --min-p 0.0 \
    --reasoning-effort medium

BASH

apt-get update
apt-get install pciutils build-essential cmake curl libcurl4-openssl-dev -y
git clone https://github.com/ggml-org/llama.cpp
cmake llama.cpp -B llama.cpp/build \
    -DBUILD_SHARED_LIBS=OFF -DGGML_CUDA=ON
cmake --build llama.cpp/build --config Release -j --clean-first --target llama-cli llama-mtmd-cli llama-server llama-gguf-split
cp llama.cpp/build/bin/llama-* llama.cpp

BASH

pip install -U "huggingface_hub[cli]"
hf download unsloth/Qwen3.8-27B-GGUF \
    --local-dir unsloth/Qwen3.8-27B-GGUF \
    --include "*UD-Q4_K_XL*" # Use "*UD-Q3_K_XL*" for 3-bit

文本

./llama.cpp/llama-cli \
    --model unsloth/Qwen3.8-27B-GGUF/Qwen3.8-27B-UD-Q4_K_XL.gguf \
    --temp 1.0 \
    --top-p 0.95 \
    --top-k 20 \
    --min-p 0.0

文本

uv venv unsloth-nvfp4-env --python 3.13
source unsloth-nvfp4-env/bin/activate
uv pip install "vllm>=0.25.0" "flashinfer-python>=0.6.13" "nvidia-cutlass-dsl>=4.5.2" \
    --torch-backend=auto

文本

vllm serve unsloth/Qwen3.8-27B-NVFP4

文本

vllm serve unsloth/Qwen3.8-27B-NVFP4
    --speculative-config '{"method": "mtp", "num_speculative_tokens": 2}'

BASH

sudo apt-get update
sudo apt-get install -y ffmpeg

文本

uv pip install sglang sglang-kernel \
  --extra-index-url https://sgl-project.github.io/whl/cu130/ \
  --extra-index-url https://download.pytorch.org/whl/cu130 \
  --index-strategy unsafe-best-match

文本

sglang serve unsloth/Qwen3.8-27B-NVFP4

文本

sglang serve unsloth/Qwen3.8-27B-NVFP4 --speculative-algorithm EAGLE \
    --speculative-num-steps 3 --speculative-eagle-topk 1  --speculative-num-draft-tokens 4

BASH

apt-get update
apt-get install pciutils build-essential cmake curl libcurl4-openssl-dev -y
git clone --branch iq1-narrow https://github.com/unslothai/llama.cpp
cmake llama.cpp -B llama.cpp/build \
    -DBUILD_SHARED_LIBS=OFF -DGGML_CUDA=ON
cmake --build llama.cpp/build --config Release -j --clean-first --target llama-cli llama-mtmd-cli llama-server llama-gguf-split
cp llama.cpp/build/bin/llama-* llama.cpp

BASH

pip install -U "huggingface_hub[cli]"
hf download unsloth/Qwen3.8-2.4T-A95B-GGUF \
    --local-dir unsloth/Qwen3.8-2.4T-A95B-GGUF \
    --include "*Q1_0*" # Use "*IQ2_XXS*" for 2-bit

文本

./llama.cpp/llama-cli \
    --model unsloth/Qwen3.8-2.4T-A95B-GGUF/UD-Q1_0/Qwen3.8-2.4T-A95B-UD-Q1_0-00001-of-00010.gguf \
    --temp 1.0 \
    --top-p 0.95 \
    --top-k 20 \
    --min-p 0.0

BASH

pip install -U "huggingface_hub[cli]"
hf download unsloth/Qwen3.8-2.4T-A95B-GGUF \
    --local-dir unsloth/Qwen3.8-2.4T-A95B-GGFF \
    --include "*IQ1_S*" # Use "*IQ2_XXS*" for 2-bit

文本

./llama.cpp/llama-cli \
    --model unsloth/Qwen3.8-2.4T-A95B-GGUF/UD-IQ1_S/Qwen3.8-2.4T-A95B-UD-IQ1_S-00001-of-00012.gguf \
    --temp 1.0 \
    --top-p 0.95 \
    --top-k 20 \
    --min-p 0.0