cli-basic.json

{
  // Minimal TensorSharp.Cli configuration.
  // Run with:  TensorSharp.Cli --config config/cli-basic.json
  // Any option you also pass on the command line overrides the value here,
  // e.g.  TensorSharp.Cli --config config/cli-basic.json --max-tokens 512
  //
  // The model auto-downloads into ${modelRoot} on the first run (real, ungated
  // GGUF, ~8.9 GB) and is reused on every later run.
  "variables": {
    "modelRoot": "${TENSORSHARP_MODELS:-../models}",
    "hf": "https://huggingface.co"
  },

  "backend": "ggml_cuda",
  "max-tokens": 256,
  "temperature": 0.7,
  "top-p": 0.9,

  "model": {
    "path": "${modelRoot}/Qwen3.5-9B-Q8_0.gguf",
    "urls": [ "${hf}/unsloth/Qwen3.5-9B-GGUF/resolve/main/Qwen3.5-9B-Q8_0.gguf" ]
  }
}