---
license: mit
language:
  - en
base_model:
  - litert-community/gemma-4-E2B-it-litert-lm
  - google/gemma-4-E2B-it
tags:
  - litert
  - litert-lm
  - on-device
  - mobile
  - gemma
  - gemma-4
  - flutter
  - healthcare
  - dementia
  - alzheimers
  - face-recognition
  - arcface
  - kaggle-gemma-4-good-hackathon
pipeline_tag: text-generation
library_name: flutter_gemma
---

# Zekra — Memory Care for Alzheimer's

> *A digital life-story book for Alzheimer's, running entirely on-device.*

**Website:** [zekra.live](https://zekra.live)
**App source:** [github.com/aelhajj/zekra-ai](https://github.com/aelhajj/zekra-ai)
**Submission:** Kaggle **Gemma 4 Good Hackathon** — Impact Track (Health & Sciences), with eligibility for the **LiteRT** and **Unsloth** Special Technology Tracks. Deadline 2026-05-18.

Zekra helps people with Alzheimer's recognize the faces around them and recall the stories that go with those faces. A caregiver builds a small graph of the family — who everyone is, how they are related, the photos and memories that matter. The patient lifts the phone, points the camera at someone, and Zekra tells the story warmly:

> *"This is your grandson Ahmad. Last summer you took him to the park and he laughed at the ducks."*

No "do you remember." No quiz. Just sharing. Everything runs on the phone — no backend, no cloud, no account, photos never leave the device.

## What's in this repo

Two on-device models powering the Zekra Flutter app:

| File | Size | Purpose |
|---|---|---|
| [`model-dyn-wi8-afp32.litertlm`](./model-dyn-wi8-afp32.litertlm) | 5.1 GB | Fine-tuned **Gemma 4 E2B** care-dialogue model, exported as a `.litertlm` bundle for **LiteRT-LM** (GPU full-delegation on Pixel 9 Pro). |
| [`arcface_zekra_r50_fp16.onnx`](./arcface_zekra_r50_fp16.onnx) | 83 MB | Fine-tuned **InsightFace ArcFace** face embedder (ResNet50, 512-dim, fp16) for cross-age + kinship-aware face verification, served via `onnxruntime`. |

Both files are dropped into the Flutter app's private documents folder at runtime (see deployment recipe below).

## Gemma 4 E2B — care-dialogue fine-tune

Fine-tuned from **[`litert-community/gemma-4-E2B-it-litert-lm`](https://huggingface.co/litert-community/gemma-4-E2B-it-litert-lm)** (which is itself a LiteRT-LM port of **[`google/gemma-4-E2B-it`](https://huggingface.co/google/gemma-4-E2B-it)**) with Unsloth LoRA + a hand-curated SFT corpus of multi-turn care dialogues. The model is wired as a **router**, not a free agent loop — it picks from twelve tools (`identify_face`, `tell_me_about`, `whos_with_me`, `where_am_i`, `check_meds`, `log_dose_taken`, `create_reminder`, `make_note`, `get_help`, etc.). Pre-routers on the Flutter side catch distress signals and face-recognition intents deterministically before the model sees the turn.

**Training**
- **Base:** Gemma 4 E2B (instruction-tuned)
- **Method:** Unsloth LoRA — all-linear, r=16, α=32 — then merged into the base for export
- **Corpus:** 3,198 hand-curated turns across 19 waves, generated by ~9 Flutter instances running in parallel on an M4 MacBook (Claude played the patient, Gemma played the companion), LLM-judged per batch
- **Doctrine:** **never quiz, just tell** — backed by Tom Kitwood's *Dementia Reconsidered* (1997) and the DAWN Method's guidance against recall-testing
- **Distress detection** uses [Al-Mosaiwi & Johnstone (2018)](https://doi.org/10.1177/2167702617747074) absolutist-word patterns
- **Clinical grounding:** [Cochrane review (Woods et al., 2018)](https://www.cochranelibrary.com/cdsr/doi/10.1002/14651858.CD001120.pub3/full) on reminiscence therapy + [DEMENTIA-PLAN (2025)](https://arxiv.org/html/2503.20950v1) on knowledge-graph-grounded dialogue

**Export**
- **Quantization:** dynamic int8 weights, fp32 activations (`wi8 afp32`) — wi4 collapsed quality, wi8 was the sweet spot that kept persona while loading on GPU
- **Format:** `.litertlm` bundle, 12-section layout, `SP_Tokenizer` spliced from the base bundle so prefill/decode token IDs stay aligned with the LiteRT-LM runtime
- **Runtime:** [`flutter_gemma`](https://pub.dev/packages/flutter_gemma) on Android/macOS, GPU backend, full delegation (865/865 + 1533/1533 nodes, single partition each), warm-up ~2.5 s on Pixel 9 Pro

The full export recipe (including the dead-ends — `--task text_generation` SEGVs in `embedding_lookup`, `--experimental_lightweight_conversion` causes `MUL(const, const)` GPU rejection, `add_hf_tokenizer` produces token salad) is on [our blog](https://zekra.live/#blog) and in the working Kaggle notebook: [Fine-Tuned Gemma 4 on Mobile — LiteRT Solution](https://www.kaggle.com/code/aelhajj/fine-tuned-gemma-4-on-mobile-lite-rt-solution/notebook).

## ArcFace — cross-age face embedder

Fine-tuned from the InsightFace **`w600k_r50`** ArcFace ResNet50 backbone (the `buffalo_l` model pack, from [insightface/python-package](https://github.com/deepinsight/insightface/tree/master/python-package)).

**Training**
- **Loss:** CosFace classification (s=15, m=0.35) — pair-contrastive collapsed; switching loss families fixed it
- **Data:** 68,000 images across 8,651 identities from six datasets — [FG-NET](http://yanweifu.github.io/FG_NET_data/) (cross-age), [Families in the Wild](https://web.northeastern.edu/smilelab/fiw/) (kinship hard-negatives), [CFP-FP](http://www.cfpw.io/) and [CPLFW](http://www.whdeng.cn/CPLFW/index.html) (frontal-vs-profile), [RMFRD](https://github.com/X-zhangyang/Real-World-Masked-Face-Dataset) and [MLFW](http://whdeng.cn/MLFW/index.html) (masked faces)
- **Tricks:** BN frozen, head warmup, full backbone unfrozen after warmup
- **Output:** 512-dim L2-normalized embeddings, fp16 ONNX

**Verification accuracy** (held-out 13.8K pairs, cos ≥ 0.40 threshold):

| Model | Accuracy | Separation |
|---|---|---|
| `w600k_r50` (stock) | 75.1% | 1.0× |
| Zekra fine-tune (this file) | **95.5%** | **2.8×** |

**Runtime:** [`onnxruntime`](https://pub.dev/packages/onnxruntime) on Android/macOS. Embedded into a 512-dim HNSW vector index in ObjectBox on the device; cosine lookup at recognition time.

## How the two models work together

```
camera frame
   ↓  Google ML Kit (on-device face detection)
face crop
   ↓  ArcFace ONNX  (this repo — 83 MB)
512-dim embedding
   ↓  ObjectBox HNSW cosine search
PersonEntity match  ──→  graph lookup (relationship, recent memories, photos)
                            ↓  prompt assembly + Gemma 4 call
                       Gemma 4 LiteRT (this repo — 5.1 GB)
                            ↓
                       warm sentence (TTS optional)
```

No network. Photos never leave the device.

## How to use

The shipping integration is the [Zekra Flutter app](https://github.com/aelhajj/zekra-ai) — `flutter_gemma` for the LLM, `onnxruntime` for ArcFace, ObjectBox for the on-device knowledge graph + vector search. Recipe to drop both files onto a Pixel:

```bash
# Bundle (5.1 GB) — push, then run-as cp into the app's private dir
adb push model-dyn-wi8-afp32.litertlm /data/local/tmp/model.litertlm
adb shell "run-as com.zekra.zekra cp /data/local/tmp/model.litertlm app_flutter/model.litertlm"

# ArcFace (83 MB)
adb push arcface_zekra_r50_fp16.onnx /data/local/tmp/w600k_r50.onnx
adb shell "chmod 0666 /data/local/tmp/w600k_r50.onnx"
adb shell "run-as com.zekra.zekra cp /data/local/tmp/w600k_r50.onnx app_flutter/w600k_r50.onnx"
```

Both files land in `/data/user/0/com.zekra.zekra/app_flutter/`. The app picks them up on the next launch.

## License

MIT (this repo). Upstream licenses apply to the base models — see the [Gemma terms](https://ai.google.dev/gemma/terms) for the LLM base (Apache 2.0 on the `litert-community` mirror, Gemma terms on the Google original) and the [InsightFace license](https://github.com/deepinsight/insightface/blob/master/LICENSE) for the ArcFace backbone.

## Citation

```
@misc{zekra2026,
  title  = {Zekra: A Digital Life Story Book for Alzheimer's, Running Entirely On-Device},
  author = {El Hajj, Amanie and El Hajj, Hadi},
  year   = {2026},
  url    = {https://zekra.live},
  note   = {Kaggle Gemma 4 Good Hackathon submission — Health \& Sciences track}
}
```

---

*Zekra (ذكرة) means **memento** in Arabic. Built for our grandmother, and one day perhaps for ourselves.*