Datasets:

ppak10
/

Agentic-SLS-Knowledge

license: mit
tags:
  - sls
  - 3d-printing
  - additive-manufacturing
  - inova-mk1
  - agents
configs:
  - config_name: knowledge
    data_files:
      - split: train
        path: data/knowledge.jsonl
    default: true

SLS-Knowledge

Knowledge corpus for SLS printing — the Inova MK1 specifically and selective laser sintering in general: technical data sheets, machine data, parameter translation notes, and (as the agentic system evolves) knowledge the agents themselves write down and maintain.

This is also the runtime knowledge base of the Inova MK1 agentic SLS system: the reference_list/reference_get MCP tools serve source/*.md directly to every agent harness. The parent repo pins this repo's commit as a submodule, so the exact corpus state agents saw at any experiment is reconstructible.

source/*.md — curated documents (markdown + title/summary frontmatter) — these are what the agents' reference_* tools serve
source/materials/<powder>/ — original manufacturer documents, one folder per powder (currently the official Formlabs TDS PDFs for all seven Fuse-series SLS powders: Nylon 12, Nylon 12 GF, Nylon 12 White, Nylon 12 Tough, Nylon 11, Nylon 11 CF, TPU 90A)
data/knowledge.jsonl — processed, load_dataset-ready form of the markdown corpus (scripts/knowledge/01_extract.py)

Current documents: Formlabs Nylon 12 GF TDS values + measured baselines (the optimization targets), Fuse 1 → Inova MK1 parameter translation, Inova print-profile conventions.