Echo voices and Baobab national models — Africa-first language intelligence.

Explore Baobab
Skip to content

Lugemi status

Baobab Amharic research slot

Generative LM research_planned surface for Amharic — no foundation weights claimed.

model_id
baobab-am
model_version
2026.10.national
phase
research
status
research_slot_shipped
as_of
2026-10-23T10:15:08.982Z

research

id
baobab-am-generative-lm
status
research_planned
foundation_llm_weights
false
generative_inference_live
false
ships_trained_baobab_foundation
false
separate_variety_checkpoints_required
true
training_recipe
docs/baobab/GENERATIVE_LM_RECIPE.md
note
Research slot + harness + training recipe only. Do not brand phrase MT or speech corridors as a full national LLM. Generative inference unsupported until gates are met.

proposed_checkpoint_ids

primary
baobab-am-addis-lm-*

prerequisites

prerequisites: 3
  • en-am-addis corridor + national API surfaces remain the product wedge
  • Mix / Fidelity / Live evals and rights clear
  • Consented parallel + speech data with held-outs

acceptance_gates

acceptance_gates: 5
  • held_out_speakers

    id
    held_out_speakers
    required
    true
    met
    false
    description
    Held-out evaluation speakers must never appear in training exports; speaker lists remain empty until consented collection.
  • community_consent

    id
    community_consent
    required
    true
    met
    false
    description
    trainingEligible stays false until Addis Ababa / Ethiopia community speakers re-confirm for training.
  • variety_separation

    id
    variety_separation
    required
    true
    met
    false
    description
    Any future LM/adapters must ship as baobab-am-addis-* — never one blob marketed as interchangeable varieties.
  • beat_phrase_mt_baselines

    id
    beat_phrase_mt_baselines
    required
    true
    met
    false
    description
    Generative claims require held-out gains vs corridor phrase-MT baselines on declared native tasks.
  • commercial_rights_clear

    id
    commercial_rights_clear
    required
    true
    met
    false
    description
    Base model + data licenses must clear commercial ministry/enterprise redistribution before promotion.

task_harness

goldens
apps/api/src/baobab/baobab-am.research.ts#RESEARCH_GOLDENS
runner
apps/api/src/baobab/baobab-am.research.ts#evaluateBaobabAmNativeTasks
generative_default
unsupported

task_eval_summary

cases
3
generative_unsupported
3
corridor_mt_baseline_usable
1
note
Harness proves generative paths are unsupported and records phrase-MT baselines only.