https://github.com/Niko1221/Strata/releases
下载:
strata-control-0.1.55-windows-x86_64.zip
strata-engine-0.1.55-windows-x86_64-cuda-cuda13.tar.gz
解压strata-control-0.1.55-windows-x86_64.zip到E:\Strata-0.2.9_windows
解压strata-engine-0.1.55-windows-x86_64-cuda-cuda13.tar.gz到E:\Strata-0.2.9_windows\engine
cd E:\Strata-0.2.9_windows
新建strata-unsloth-ud-q4_k_xl.json
{
"exe": "E:\\Strata-0.2.9_windows\\engine\\strata.exe",
"args": [
"--pack",
"E:\\Strata-data\\packs\\unsloth-ud-q4_k_xl",
"--native",
"E:\\Qwen3.8-Flash-Next-GGUF-UD-Q4_K_XL_unsloth\\Qwen3.8-Flash-Next-UD-Q4_K_XL-00001-of-00004.gguf",
"--expert-profile",
"E:\\Strata-0.2.9_windows\\data\\expert-profile.bin",
"--expert-cache",
"auto",
"--prefill",
"auto",
"--spec",
"4",
"--mtp",
"E:\\Strata-data\\mtp\\rt",
"--max-context",
"262144",
"--kv",
"int8",
"--kv-resident",
"32768",
"--vision",
"--pcie-frac",
"0.20",
"--spec-min-p",
"0.70",
"--adapt-every",
"1",
"--adapt-swaps",
"160",
"--adapt-decay",
"0.97"
],
"cwd": "E:\\Strata-0.2.9_windows",
"tokenizer": "E:\\Strata-data\\packs\\unsloth-ud-q4_k_xl/tokenizer",
"model_name": "qwen3.8-flash-next-unsloth-ud-q4_k_xl",
"log": "E:\\Strata-0.2.9_windows\\strata-unsloth-ud-q4_k_xl.log",
"lib_dirs": [
"D:\\Program Files\\NVIDIA GPU Computing Toolkit\\CUDA\\v13.4\\bin",
"D:\\Program Files\\NVIDIA GPU Computing Toolkit\\CUDA\\v13.4\\bin\\x64"
],
"port": 8099,
"env": {
"STRATA_HC_Q8": "1",
"STRATA_ROUTE_TAIL_SKIP": "0"
},
"gpu": 0,
"gpus_asked": true,
"api_key": "74d2c57145e6ff11d1d2faa79594477f9bc903a611af1fb20218189fbbb77d82",
"open_browser": false,
"host": "0.0.0.0",
"fit_max_tokens": true,
"reasoning_effort": "medium",
"reasoning_budget_tokens": 16000,
"vision": {
"exe": "E:\\Strata-0.2.9_windows\\engine\\strata-vision.exe",
"mmproj": "E:\\Strata-data\\models\\mmproj-Qwen3.8-Flash-Next-BF16.gguf",
"model": "E:\\Qwen3.8-Flash-Next-GGUF-UD-Q4_K_XL_unsloth\\Qwen3.8-Flash-Next-UD-Q4_K_XL-00001-of-00004.gguf",
"gpu": false,
"max_tokens": 1024,
"threads": 8
},
"args_written": [
"--pack",
"E:\\Strata-data\\packs\\unsloth-ud-q4_k_xl",
"--native",
"E:\\Qwen3.8-Flash-Next-GGUF-UD-Q4_K_XL_unsloth\\Qwen3.8-Flash-Next-UD-Q4_K_XL-00001-of-00004.gguf",
"--expert-profile",
"E:\\Strata-0.2.9_windows\\data\\expert-profile.bin",
"--expert-cache",
"auto",
"--prefill",
"auto",
"--spec",
"4",
"--mtp",
"E:\\Strata-data\\mtp\\rt",
"--max-context",
"262144",
"--kv",
"int8",
"--kv-resident",
"32768",
"--vision",
"--pcie-frac",
"0.20",
"--spec-min-p",
"0.70",
"--adapt-every",
"1",
"--adapt-swaps",
"160",
"--adapt-decay",
"0.97"
]
}
python -m serve.server --engine strata --config strata-unsloth-ud-q4_k_xl.json --port 8099