training-origin.json

{
  "method": "full",
  "base_model": "vllm-sr/Vela-1.0-Encoder-307M",
  "base_revision": "720ab37904ce15054068d429381bfae7549f00f6"
}