From 60eeb4d0ea0d1b0e83282420514fc1dee07ca0ab Mon Sep 17 00:00:00 2001 From: tocmo0nlord Date: Sun, 26 Apr 2026 01:50:01 +0000 Subject: [PATCH] feat(remote): default synthetic-data-kit config --- packaging/remote/config.yaml | 31 +++++++++++++++++++++++++++++++ 1 file changed, 31 insertions(+) create mode 100644 packaging/remote/config.yaml diff --git a/packaging/remote/config.yaml b/packaging/remote/config.yaml new file mode 100644 index 0000000..16ae396 --- /dev/null +++ b/packaging/remote/config.yaml @@ -0,0 +1,31 @@ +# synthetic-data-kit default config +# Generated by llm-trainer bootstrap. Edit via the Config tab in the dashboard. + +paths: + input: /opt/synthetic/synthetic-data-kit/data/input + parsed: /opt/synthetic/synthetic-data-kit/data/parsed + generated: /opt/synthetic/synthetic-data-kit/data/generated + curated: /opt/synthetic/synthetic-data-kit/data/curated + final: /opt/synthetic/synthetic-data-kit/data/final + +llm: + provider: ollama + api_base: http://localhost:11434 + model: llama3.1:8b + temperature: 0.3 + max_tokens: 2048 + +ingest: + chunk_size: 2000 + chunk_overlap: 200 + +create: + default_pair_type: qa + default_num_pairs: 50 + +curate: + threshold: 7.0 + use_llm_judge: true + +save: + default_format: jsonl \ No newline at end of file