ruos-foundry-cu-qwen3-30b-a3b-e64

ruos-foundry-cu-qwen3-30b-a3b-e64 is a ruOS specialist cut from Qwen/Qwen3-30B-A3B-Instruct-2507 with MoE-Foundry (ADR-064): the parent's routed experts were traced on the ruOS ruos-cu calibration set, the 64 most-used experts per routed layer were kept, and a complete smaller Mixture-of-Experts model was exported — same backbone, same tokenizer, same token top-k, fewer experts per layer. No weight was changed; experts were removed and renumbered and the router rows sliced to match.

Status: unevaluated and disabled

Per MoE-Foundry's rule an exported specialist starts quality_status: unevaluated, enabled: false. A structural export proves tensor integrity, not retained capability. This checkpoint is published so that slim-eval (slim/eval: run, redteam, regression, verdict) can measure it against the parent on the frozen ruOS test split; until that verdict is recorded here it must not be routed to. The full parent remains the fallback.

Parent


repo	`Qwen/Qwen3-30B-A3B-Instruct-2507`
revision	`0d7cf23991f47feeb3a57ecb4c9cee8ea4a17bfe`
licence	Apache-2.0
MoE-Foundry parent_id	`44c2b6971fe6979409ac72586db5fa4abfd73746afc52607368a59740848149d`
experts per routed layer (parent → this)	128 → 64
routed layers	48
token top-k (unchanged)	8

Calibration

Domain ruos-cu: 148 rows (148 from the val split, 0 topped up from train; the test split is never used for calibration), ~63523 tokens, families {"browser_step":142,"cu_step":6}, licences {"project-owned":148}. Calibration file sha256 94a5527b141269ab5078a14e3d129b387aa69412c65334160715fd9a58dbfef5. Texts are the ChatML prompt plus the reference answer.

Router traces: 412 tasks, 148144 tokens, 7110912 rows on NVIDIA A100-SXM4-80GB (bfloat16, transformers 4.51.3); trace sha256 7aabb05cbf715156437393fc5a5b0d756951ecee35eb4542a06fc812f11f9403.

Selection method: mass (accumulated routing probability per expert per layer) — a usage proxy, not causal importance.

Receipt


specialist_id	`1409f5fef6a8bfbef4be7f1f2e91700f7c3e7d15690ec8a01a37ed2a1c6ca1ea`
checkpoint_id	`5a74281b02e586e0c6a9e79323f61c90884e188c6b4719fd91ad600b94d2c0da`
mask_sha256	`d8ea421921a28f6a9c4f0ae3f1bc01efc159fad83163d73c8a2629a3f0b9d721`
parent_id	`44c2b6971fe6979409ac72586db5fa4abfd73746afc52607368a59740848149d`
input tensor bytes	61064245248
output tensor bytes	32060633088
reduction	47.5%

separator_receipt.json in this repo is the full MoE-Foundry receipt including the per-layer expert mask.

Files

path	bytes	sha256
LICENSE	11343	05cab46843576551502bfdf712f84e93e6e9590d9997306ed4f6635ef82811d9
SHA256SUMS	737	7761e07f29ce824f3893518837e285cc4d1a4c60ba1f3e295d00945c7618a10a
config.json	963	f0bf60e7da89b14b5763344441e4f60eccda09f62c0111c5f178283c5f3d81ec
generation_config.json	239	19d306dd769db12a9d710b44cf7f83b635efbe5166b84fb4358a08fb7d88bb53
merges.txt	1671839	599bab54075088774b1733fde865d5bd747cbcc7a547c5bc12610e874e26f5e3
model.safetensors	32061839304	54ae831f3df008547d5734fab1a2bd3a50349f2dbcb481ad418a57e318a18742
separator_receipt.json	32964	dd6ec404f4400ee8e7e26cb40a2d6e4c3581c4c165232042ff89c379f17b9b56
tokenizer.json	11422654	aeb13307a71acd8fe81861d94ad54ab689df773318809eed3cbe794b4492dae4
tokenizer_config.json	9377	a62ff0a2472a0fa1b8eaabcb57c59b58afa42a22831dc141400b6e0cf2b65ce3
vocab.json	2776833	ca10d7e9fb3ed18575dd1e277a2579c16d108e32f27439684afa0e10b1440910

Run

vllm serve ruvnet/ruos-foundry-cu-qwen3-30b-a3b-e64 --max-model-len 8192

Loads with transformers as a standard qwen3_moe checkpoint (single safetensors file, num_experts reduced in config.json).

Limitations

Unevaluated: no capability, memory or latency claim is made here.
Retained experts were chosen by routing mass on ruOS calibration prompts; requests outside that domain should go to the parent.
Memory: fewer experts means a smaller checkpoint; loading several specialists next to the parent can use more total memory than the parent alone.

Provenance

MoE-Foundry 6677a25 (moe-separator inspect → profile_hf → select → export → mixture)
run foundry-20260907T172935Z-qwen3-30b-a3b on a single vast.ai GPU; ruos-desktop slim/foundry + slim/scripts/foundry-e2e.sh
authorisation: rUv, "implement this using ruvnet/MoE-Foundry using vast.ai in a worktree, implement e2e and push models to repo"