vrr/BigMoeOnEdge
2
0
Fork 0
mirror of https://github.com/Helldez/BigMoeOnEdge.git synced 2026-10-03 03:25:42 +00:00
Code Issues Projects Releases Packages Wiki Activity Actions

docs(android): list Qwen3.6 in the catalog, and separate sweep from default

Browse source 
The app README enumerated a two-model catalog; Qwen3.6-35B-A3B has shipped
in it since 0.11.0. Qwen3-30B is labelled to match what the app displays,
the release-signing prerequisite is stated so a published build is
reproducible, and the 4000 MiB in the expected-numbers section is marked as
a sweep point rather than the shipped default, which is a fixed 2000 MiB.

prefetch.md and the config.h comment both said prefetch needs
--cache-mb > 0; validate() also accepts --cache-mb auto.
This commit is contained in:
Helldez 2026-07-19 11:27:44 +02:00
parent d10d10f1b8
commit 022758e467
3 changed files with 16 additions and 8 deletions
Download patch file Download diff file Expand all files Collapse all files

14
examples/android/README.md
View file

@ -28,6 +28,11 @@ research harness and keeps the app a thin driver over the CLI.
adb install app/build/outputs/apk/dev/debug/app-dev-debug.apk
```
Published sideload builds are release-signed with a stable key instead, so an update installs
over the previous one rather than being refused. That needs a `keystore.properties` next to
`app/` (gitignored — it points at the keystore and holds its passwords); without it the release
build falls back to debug signing.
## Flavors
Two build flavors differ only in how a model reaches the device:
@ -43,7 +48,8 @@ The picker lists every MoE `.gguf` it finds (dense models are filtered out by a
check). Nothing below needs a storage permission except the last option.
1. **Built-in catalog** (both flavors) — the "Get a model" card offers the models this engine
is measured on, each a single tap: **Qwen3-30B-A3B-Q4_K_M** (~18.5 GB, the reference model)
is measured on, each a single tap: **Qwen3-30B-A3B-Q4_K_M** (~18.6 GB, the reference model),
**Qwen3.6-35B-A3B-Q4_K_M** (~22.3 GB, a hybrid attention/SSM MoE, comfortably past device RAM)
and **Gemma-4-26B-A4B-it-Q4_K_M** (~17 GB). Downloads run in a foreground worker, survive the
app being killed, resume an interrupted transfer instead of restarting, and appear in the
picker when done.
@ -84,5 +90,7 @@ adb push gpt-oss-120b-Q4_K_M.gguf /data/local/tmp/bmoe/ # or import it with th
On a phone with UFS 4.x storage and ~12 GB RAM, streaming Qwen3-30B-A3B-Q4_K_M with the
expert cache at 4000 MiB, 4 I/O lanes and 4 compute threads, decode settles around
**0.55–0.6 s/token (~1.8 tok/s)** — a model ~1.7× the device RAM, lossless. See
`../../docs/benchmark-method.md` for the full procedure and the cache/thread sweep.
**0.55–0.6 s/token (~1.8 tok/s)** — a model ~1.7× the device RAM, lossless. That 4000 MiB is a
sweep point from the benchmark protocol, not the app default: the app ships a fixed 2000 MiB
expert cache. See `../../docs/benchmark-method.md` for the full procedure and the cache/thread
sweep.