Jinghpaw · corpus in collection

Jinghpaw · ≈1.5M speakers · fine-tuning on a community-curated corpus.

I COULD BE BOUNDED IN A NUTSHELL AND COUNT MYSELF A KING OF INFINITE SPACE. [ 2.2.254-255 ] · Kachin communities, northern Myanmar & diaspora

The same thought, double the cost.

ENGLISH · EN-US
The river remembers what the map forgets.
words
8
tokens
8
tok/word
1.0
JINGHPAW · JINGHPAW
Hka gaw sumla nnan malap kau ai lam ni hpe matsing da ai.
words
12
tokens
18
tok/word
1.5
1.0 tok/wordvs1.5 tok/word· 2.3× the cost to think in your own language

Latin orthography with tone marks — vocabularies fragment tone-bearing syllables.

A Latin script, still under-tokenized

Jinghpaw is written in a Latin-based orthography developed in the late 19th century. It looks familiar to a tokenizer trained on English but the diacritics, tone marks, and syllable structure fragment into pieces the vocabulary never learned to keep together.

The corpus is in collection with community and church archives. Provenance is per-item, consent is revocable.

section numbers · toggle blueprint for configs
[00] · plate

We work with open source models. Gemma-4 (and SEA-LION )is being used for Burmese, Karen, and Jinghpaw as well as; Llama-3.1-8B may be used for Rohingya, for byte-level fallback where Hanifi is unrepresented.

Curation
[01] · plate

Community participation and consent for curation and capture. Provenance recorded and fair compensation provided.

Normalization
[02] · plate

Converting legacy encodings and non-standard scripts into universal digital formats (e.g., Zawgyi to Unicode) is an act of cultural and historic preservation. Working alongside local linguists and community members, we systematically standardize text to be machine-readable while ensuring cultural context, orthographic nuances, and historical integrity remain intact.

[03] · plate

Adapting foundational base models to local linguistic realities through targeted, data-efficient techniques calibrated to each language’s structure and context. Training logs and loss curves are published directly to the public archive with complete technical transparency.

[04] · plate

Community-annotated benchmarks. Evals and benchmarks designed with local community and experts. Annotators acknowledgements and compensation. Inter-annotator agreement reported alongside every score.

THE CONTESTED QUESTIONS · KEPT VISIBLE, NOT RESOLVED FOR THE COMMUNITY
  • Q01How to represent tone marks consistently across religious and secular writing.
  • Q02Whether to include neighboring Kachin languages in a shared corpus or keep them distinct.
[ 06 ] THE VOICES
named with consent · faces omitted by choice
anonymous
anonymous
teacher, Myitkyina
duration · 1:32
[ 06 ] SOVEREIGNTY

The corpus belongs to the communities that made it. Consent is per item and revocable. Revocation is honored in the weights, not just in the archive.

Model weights reflect the terms and governance frameworks chosen with the community.

[ 08 ] COLOPHON

Completed at TBD, in 2026, on the Jinghpaw project, rank TBD. Released with community consent; evaluation held.

place
TBD
date
2026 — provisional
hardware
TBD
license
TBD — community-elected
elsinore data co · undertext · v0.5 · 2026.07
next projectRohingya
contribute