aether-v7.2-unified

Corrected 12 Nov 2026: an earlier version said a fleet of public GPUs trained this model and that rounds were paid. The rounds ran on company-operated machines; no round was paid and no outside contributor has trained yet.

The Sephirot adapter served by qbc.network, and the first adapter produced by our decentralised training loop to pass a full evaluation and be promoted into the live served model. It succeeds aether-v7.1-unified.

It is a small adapter over a frozen Qwen2.5-7B-Instruct base (Q4_K_M): ten low-rank experts with a top-2 router, about 1.18 million trainable parameters. The base model is not redistributed here; load Qwen/Qwen2.5-7B-Instruct and apply this adapter on top. Serving precision is BF16.

On-chain attestation

This model's identity is attested on the QBC chain. The on-chain weight_root is a SHA-256 over (base, adapter, tokenizer, config):

  • weight_root: 0x3e99b3d2f4e4d5eacbb3fe8d7e2f709dbfc48deb34407035383b3a5263e29e4e
  • registry round: 2, finalised by a 4-of-5 validator quorum
  • submitted at block: 1698355
  • content root of the trained adapter: 0x157a01b9

How it was trained

Trained with a DiLoCo-style local-training-plus-aggregation loop on company-operated machines. No round was paid and no outside contributor has trained yet. A vocab-chunked backward pass keeps a full 7B adapter step inside roughly 7.5GB of video memory, so commodity cards can train the served model. Training runs as short rounds of about 35 minutes: a worker draws a fresh shuffled batch from a large corpus, trains the adapter locally, and submits a candidate. A candidate is accepted only if it beats the previous accepted candidate on a fixed held-out set it never trains on. Ten rounds produced seven accepted units, and the best by full held-out evaluation is this release. All rounds ran on company-operated machines and none was paid.

Evaluation

Measured on this adapter over the frozen base, as a snapshot from the promotion evaluation (not a live probe):

Measurement Base This adapter Change
Held-out cross-entropy (clean 1000-example corpus, nats) 2.9014 2.5609 -0.3405
MMLU (full 14042, uncapped) 0.7116 0.7109 -0.07pp (flat)
GSM8K (full 1319, uncapped) 0.7248 0.7498 +2.50pp

The BF16 weights served here reproduce the held-out CE of the F32 candidate to four decimals, so the -0.3405 nats generalisation gain holds at serving precision. Generalisation on unseen text improved, general knowledge held flat, and multi-step maths reasoning improved, at no measured cost to base capability. Against the previously-served adapter, the promotion gate measured a further held-out CE improvement (2.5912 to 2.4811 on the gate's own holdout).

Scope and honesty

This adapter is the product of our decentralised training mechanism (built), run on our own machines, with an owner-gated promotion into the served model. This is Stage 1 of a four-stage plan. It is not yet a permissionless model update: that stage is designed, not built, and is gated on robust aggregation that survives a malicious contributor, a stake-and-slash economic layer, and honest independent re-derivation of the claimed improvement.

During this work our own evaluation flattered us three times, through memorisation on training data, a sampling-window bug, and a mixed-scale gate that false-rejected a good round. Each was caught and fixed before any number here was trusted. The full account is in the engineering write-up.

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel

base = AutoModelForCausalLM.from_pretrained("Qwen/Qwen2.5-7B-Instruct")
tok = AutoTokenizer.from_pretrained("Qwen/Qwen2.5-7B-Instruct")
model = PeftModel.from_pretrained(base, "QuantumAI-Blockchain/aether-v7.2-unified")

Licence

Released under Apache-2.0, matching the Qwen2.5-7B-Instruct base licence.

Downloads last month
-
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for QuantumAI-Blockchain/aether-v7.2-unified

Base model

Qwen/Qwen2.5-7B
Adapter
(2,934)
this model

Evaluation results

  • accuracy on MMLU (full, 14042 questions, uncapped)
    self-reported
    0.711
  • accuracy on GSM8K (full, 1319 questions, uncapped)
    self-reported
    0.750