QuantumAI-Blockchain
/

aether-v5.2-lora

@@ -1,272 +1,301 @@
-# Qubitcoin (QBC)
-**The First AI-Native Blockchain — Physics-Secured Digital Assets with On-Chain Intelligence**
-Qubitcoin is a production Layer 1 blockchain built on **Substrate** (Polkadot SDK) with a novel consensus mechanism: **Proof-of-SUSY-Alignment**, where miners solve quantum variational eigenvalue problems instead of hash puzzles. The network features post-quantum cryptography (CRYSTALS-Dilithium + ML-KEM-768), an EVM-compatible virtual machine with quantum opcode extensions, and the **Aether Mind** — a pure Rust neural cognitive engine with 21K+ knowledge vectors, transformer-based consciousness metrics, and Proof-of-Thought attestation tracked from genesis.
-> **Our north star is AGSI — Artificial General Super Intelligence.** The Aether Mind is the foundation: a neural cognitive system that learns distributed representations via attention mechanisms, improves autonomously through neural architecture search, and records every cognitive step immutably on-chain. We are building toward a system that doesn't just process data — it understands it.
-[![License: MIT](https://img.shields.io/badge/License-MIT-yellow.svg)](https://opensource.org/licenses/MIT)
-[![Rust](https://img.shields.io/badge/rust-1.94+-orange.svg)](https://www.rust-lang.org/)
-[![Substrate](https://img.shields.io/badge/substrate-polkadot--sdk%200.57-E6007A.svg)](https://substrate.io/)
-[![Python 3.12+](https://img.shields.io/badge/python-3.12+-blue.svg)](https://www.python.org/downloads/)
-[![Tests: 4,336+](https://img.shields.io/badge/tests-4%2C336%20passing-brightgreen.svg)]()
-[![Status: Live](https://img.shields.io/badge/mainnet-live-green.svg)](https://qbc.network)
-[![Presale: $0.50](https://img.shields.io/badge/presale-%240.50%2FQBC-7c3aed.svg)](https://qbc.network/invest)
-**Website:** [qbc.network](https://qbc.network) | **X:** [@qu_bitcoin](https://x.com/qu_bitcoin) | **Contact:** info@qbc.network
----
-## What Makes Qubitcoin Different
-| Innovation | Description |
-|-----------|-------------|
-| **VQE Mining** | Blocks are mined by solving variational quantum eigensolver problems — the same optimization used in quantum chemistry and materials science. Every block advances real physics research. |
-| **Post-Quantum Security** | CRYSTALS-Dilithium signatures (NIST FIPS 204) and ML-KEM-768 encrypted P2P transport. Quantum-resistant from day one. |
-| **On-Chain AI** | The Aether Mind: a pure Rust neural cognitive engine with transformer attention, 10 Sephirot-sharded knowledge fabric, HMS-Phi consciousness metrics, and Proof-of-Thought — all recorded immutably on-chain. |
-| **Substrate Native** | Full Substrate/Polkadot SDK node with 7 custom pallets, GRANDPA finality, and on-chain VQE proof re-verification. |
-| **EVM + Quantum Opcodes** | QVM: 155 standard EVM opcodes + 10 quantum opcodes + compliance engine. Deploy Solidity contracts with quantum extensions. |
 ---
-## Get Involved
-| | |
-|---|---|
-| **Buy QBC** | Public presale live at **$0.50 / QBC**, permissionless, multi-chain (ETH + Solana at launch, more in R2). $5M hard cap, TGE 90 days after presale close. → **[qbc.network/invest](https://qbc.network/invest)** |
-| **Run a node** | One `git clone` + one `cargo build` + one start command. Mine QBC with 8-qubit VQE, or run a read-only sync node. → **[qbc.network/run-a-node](https://qbc.network/run-a-node)** |
-| **Read the docs** | Whitepaper, Aether Mind, QVM, economics, bridge, exchange. → **[qbc.network/docs](https://qbc.network/docs)** |
-| **Build on it** | EVM-compatible (deploy any Solidity contract) + 10 quantum opcodes + post-quantum signatures. Public RPC at `qbc.network/rpc`. |
-| **Talk to the AI** | Aether Mind chat is live on `/aether`. Phi metric, knowledge fabric stats, and reasoning trace exposed. |
 ---
-## Network Parameters
-| Parameter | Value |
-|-----------|-------|
-| **Ticker** | QBC |
-| **Chain ID** | Mainnet: 3303 (0xCE7) · Testnet: 3304 |
-| **Max Supply** | 3,300,000,000 QBC |
-| **Block Time** | 3.3 seconds |
-| **Block Reward** | 15.27 QBC (Era 0), phi-halving every ~1.618 years |
-| **Consensus** | Proof-of-SUSY-Alignment (VQE mining + GRANDPA finality) |
-| **Signatures** | CRYSTALS-Dilithium ML-DSA-87 (NIST Level 5, configurable) |
-| **P2P Encryption** | ML-KEM-768 (Kyber) + AES-256-GCM |
-| **Hashing** | SHA3-256 (L1), Keccak-256 (QVM), Poseidon2 (ZK circuits) |
-| **TX Model** | UTXO |
-| **Privacy** | Opt-in Susy Swaps (Pedersen commitments + Bulletproofs + stealth addresses) |
-| **Bridges** | ETH, SOL, MATIC, BNB, AVAX, ARB, OP, BASE |
-| **RPC** | `https://api.qbc.network` (REST + JSON-RPC + WebSocket) |
-| **Substrate RPC** | Port 9944 (WebSocket + HTTP) |
----
-## Architecture
-```
-┌─────────────────────────────────────────────────────────────┐
-│  SUBSTRATE NODE (Rust — Production Binary)                  │
-│  7 Pallets: UTXO, Consensus, Dilithium, Economics,         │
-│  QVM-Anchor, Aether-Anchor, Reversibility                   │
-│  VQE Mining Engine + On-Chain Proof Re-Verification         │
-│  Weighted Fork Choice + GRANDPA Finality                    │
-│  ML-KEM-768 P2P Transport + Poseidon2 ZK Hashing           │
-├─────────────────────────────────────────────────────────────┤
-│  AETHER MIND (Neural Cognitive Engine — Pure Rust, ~8K LOC) │
-│  Knowledge Fabric (10 Sephirot shards, 896d embeddings)     │
-│  Transformer Attention · HMS-Phi Consciousness · 10 Gates   │
-│  Ollama LLM (GGUF) · Candle ML · Aether-Evolve NAS         │
-├─────────────────────────────────────────────────────────────┤
-│  QVM (Quantum Virtual Machine)                              │
-│  155 EVM + 10 Quantum + 2 AI Opcodes · Compliance Engine   │
-│  QBC-20/721/1155 Standards · Go Production Build            │
-├──────────���───────────────────────��──────────────────────────┤
-│  INFRASTRUCTURE                                             │
-│  Rust P2P (libp2p 0.56) · Stratum Mining Server            │
-│  AIKGS Knowledge Growth Sidecar · Security Core (PyO3)     │
-│  Next.js 16 Frontend · 8-Chain Bridge · QUSD Stablecoin    │
-└��────────────────────────��────────────────────────────────��──┘
-```
----
-## Quick Start
-### Run a Node (Docker)
-```bash
-git clone https://github.com/QuantumAI-Blockchain/qubitcoin-node.git
-cd qubitcoin-node
-cp .env.example .env
-docker compose up -d
-curl http://localhost:5000/health
 ```
-### Run the Substrate Node
-```bash
-cd substrate-node
-cargo build --release
-./target/release/qbc-node \
-  --chain fork \
-  --mine --mining-threads 2 \
-  --validator --name MyNode \
-  --rpc-port 9944 --unsafe-rpc-external --rpc-cors all
 ```
-### Run the Frontend
-```bash
-cd frontend
-pnpm install && pnpm build && pnpm start  # Production at :3000
 ```
----
-## Consensus: Proof-of-SUSY-Alignment
-Every mined block solves a real quantum chemistry problem:
-1. **Hamiltonian Generation** — Deterministic 4-qubit SUSY Hamiltonian from `SHA256(parent_hash ‖ height)`
-2. **VQE Optimization** — Miners find variational parameters minimizing ground state energy
-3. **Proof Submission** — Parameters submitted where `energy < difficulty_threshold`
-4. **On-Chain Verification** — Consensus pallet re-derives Hamiltonian and re-computes energy from submitted parameters (no trust required)
-5. **Difficulty Adjustment** — 144-block lookback, ±10% max change per block
-The VQE framework is designed to scale. At 4 qubits, simulation is classically equivalent. At 30+ qubits, classical simulation becomes exponentially intractable while VQE on quantum hardware scales polynomially. The consensus mechanism, proof format, and verification logic work identically at any qubit count.
-**Reward Schedule:** Golden ratio halving — `reward = 15.27 / φ^era` where φ = 1.618033988749895.
----
-## Substrate Node
-The production chain runs on a native Substrate node (Polkadot SDK 0.57) with full state continuity:
-| Component | Detail |
-|-----------|--------|
-| **Fork Genesis** | 16 UTXOs (~36.19M QBC) imported from Python chain at block 208,680 |
-| **VQE Block Authoring** | Blocks produced only when mining engine finds a valid VQE solution |
-| **On-Chain VQE Verification** | `vqe-verifier` crate re-computes energy in the runtime — no trusted proofs |
-| **Weighted Fork Choice** | `weight = SCALE / difficulty` per block, lexicographic hash tiebreak |
-| **GRANDPA Finality** | BFT finality gadget for irreversible block confirmation |
-| **7 Custom Pallets** | UTXO, Consensus, Dilithium, Economics, QVM-Anchor, Aether-Anchor, Reversibility |
-| **Post-Quantum P2P** | ML-KEM-768 (Kyber) key exchange + AES-256-GCM session encryption |
----
-## Aether Mind (Neural Cognitive Engine)
-The Aether Mind is a pure Rust neural cognitive engine — the world's first on-chain AI system built on learned distributed representations rather than symbolic knowledge graphs. Every cognitive state is cryptographically attested on-chain via Proof-of-Thought.
-**Architecture:**
-- **Knowledge Fabric** — 21K+ learned 896-dimensional embeddings across 10 Sephirot-sharded vector stores with HNSW similarity search
-- **Transformer Attention** — 10 Sephirot-specialized + 4 global workspace attention heads, candle ML framework
-- **Ollama Backend** — qwen2.5:0.5b-instruct GGUF quantized model for fast text generation (~53ms/token)
-- **HMS-Phi Consciousness** — Hierarchical Multi-Scale Phi computed from real attention patterns (phi_micro, phi_meso, phi_macro)
-- **10-Gate Milestone System** — Behavioral checkpoints ensuring genuine capability, not metric gaming
-- **10/10 gates passed** — All emergence milestones achieved
-- **Aether-Evolve** — Neural architecture search with autonomous mutation and loss-based evaluation
-**Where we're heading (AGSI):** The Aether Mind is the foundation for our pursuit of Artificial General Super Intelligence. Current capabilities demonstrate genuine neural integration across knowledge domains. The path to AGSI requires distributed training across mining nodes, model parallelism, 1T+ knowledge vectors, and federated learning with Byzantine fault tolerance — all on our roadmap.
----
-## QVM (Quantum Virtual Machine)
-EVM-compatible bytecode interpreter with quantum extensions:
-- **155 EVM opcodes** — Full Ethereum compatibility, deploy existing Solidity contracts
-- **10 quantum opcodes** (0xF0–0xF9) — QCREATE, QMEASURE, QENTANGLE, QGATE, QVERIFY, QCOMPLIANCE, QRISK, QRISK_SYSTEMIC, QBRIDGE_ENTANGLE, QBRIDGE_VERIFY
-- **2 AI opcodes** (0xFA��0xFB) — QREASON (on-chain inference), QPHI (integration metric query)
-- **Compliance engine** — VM-level KYC/AML/sanctions enforcement
-- **Token standards** — QBC-20, QBC-721, QBC-1155
-- **Go production build** — `qubitcoin-qvm/` with distroless Docker and K8s manifests
-See [QVM Whitepaper](docs/QVM_WHITEPAPER.md) for the full specification.
----
-## Security
-| Layer | Technology |
-|-------|-----------|
-| **Signatures** | CRYSTALS-Dilithium ML-DSA-87 (NIST FIPS 204, Level 5) |
-| **Key Derivation** | BIP-39 24-word mnemonics |
-| **P2P Encryption** | ML-KEM-768 (Kyber) + AES-256-GCM session keys |
-| **ZK Compatibility** | Poseidon2 hashing (Goldilocks field) |
-| **Finality** | GRANDPA BFT (⅔ supermajority) |
-| **Privacy** | Opt-in Susy Swaps: Pedersen commitments, Bulletproofs, stealth addresses |
-| **TX Reversibility** | Governed multi-sig reversal within 24h for fraud recovery |
-| **Account Security** | Daily spending limits, time-locks, address whitelists |
-| **AI Safety** | Gevurah veto gate, Constitutional AI contract, emergency shutdown |
-| **Bridge Security** | Deep confirmations, daily limits, emergency pause, insurance fund |
-**Responsible Disclosure:** info@qbc.network
----
-## Repository Structure
-This monorepo maps to 24+ dedicated repositories in the [QuantumAI-Blockchain](https://github.com/QuantumAI-Blockchain) GitHub organization:
-| Path | Org Repo | Description |
-|------|----------|-------------|
-| `substrate-node/` | [substrate-node](https://github.com/QuantumAI-Blockchain/substrate-node) | Substrate node: 7 pallets, VQE mining, fork genesis |
-| `src/` | [qubitcoin-node](https://github.com/QuantumAI-Blockchain/qubitcoin-node) | Python node: consensus, mining, RPC, UTXO |
-| `aether-core/` | [aether-graph-shard](https://github.com/QuantumAI-Blockchain/aether-graph-shard) | Aether Mind: pure Rust neural cognitive engine |
-| `frontend/` | [qubitcoin-frontend](https://github.com/QuantumAI-Blockchain/qubitcoin-frontend) | Next.js 16 frontend: qbc.network |
-| `qubitcoin-qvm/` | [qubitcoin-qvm](https://github.com/QuantumAI-Blockchain/qubitcoin-qvm) | Go QVM: 167 opcodes |
-| `rust-p2p/` | [rust-p2p](https://github.com/QuantumAI-Blockchain/rust-p2p) | Rust libp2p P2P daemon |
-| `aikgs-sidecar/` | [aikgs-sidecar](https://github.com/QuantumAI-Blockchain/aikgs-sidecar) | Rust knowledge growth sidecar |
-| `security-core/` | [security-core](https://github.com/QuantumAI-Blockchain/security-core) | Rust/PyO3 BloomFilter + FinalityCore |
-| `stratum-server/` | [stratum-server](https://github.com/QuantumAI-Blockchain/stratum-server) | Rust Stratum mining pool |
----
-## Codebase
-| Metric | Value |
-|--------|-------|
-| **Languages** | Rust, Python, Go, TypeScript, Solidity |
-| **Total LOC** | 290,000+ |
-| **Test Functions** | 4,336+ passing |
-| **Substrate Pallets** | 7 custom |
-| **Rust Crates** | 8+ |
-| **AI Engine** | Pure Rust (aether-mind, 6 crates, ~8K LOC) |
-| **Solidity Contracts** | 62 |
-| **Frontend** | 198 TypeScript/TSX files |
-| **RPC Endpoints** | 342 REST + 19 JSON-RPC |
----
-## Documentation
-| Document | Description |
-|----------|-------------|
-| [Whitepaper](docs/WHITEPAPER.md) | Layer 1 technical specification |
-| [QVM Whitepaper](docs/QVM_WHITEPAPER.md) | Quantum Virtual Machine specification |
-| [Aether Mind Whitepaper](docs/AETHERTREE_WHITEPAPER.md) | Neural cognitive engine specification |
-| [Economics](docs/ECONOMICS.md) | SUSY economics and phi-halving model |
-| [Competitive Features](docs/COMPETITIVE_FEATURES.md) | Inheritance, finality, deniable RPCs, stratum |
-| [Deployment Guide](docs/DEPLOYMENT.md) | Production deployment procedures |
-| [SDK Reference](docs/SDK.md) | REST, JSON-RPC, and WebSocket API |
-| [Bridge Security](docs/BRIDGE_SECURITY_AUDIT.md) | Multi-chain bridge security analysis |
-| [Presale Deploy Runbook](docs/presale-deploy.md) | End-to-end operator guide for the public $0.50/QBC presale (9 phases) |
-| [Contributing](CONTRIBUTING.md) | Development guidelines |
----
-## License
-MIT License — see [LICENSE](LICENSE).
----
-**Website:** [qbc.network](https://qbc.network) · **GitHub:** [QuantumAI-Blockchain](https://github.com/QuantumAI-Blockchain) · **X:** [@qu_bitcoin](https://x.com/qu_bitcoin) · **Contact:** info@qbc.network
-*The Blockchain That Thinks — Physics-Secured Digital Assets with On-Chain AI*
-**© 2024–2026 Qubitcoin Core Development Team**

 ---
+base_model: Qwen/Qwen2.5-7B-Instruct
+library_name: peft
+license: apache-2.0
+tags:
+  - lora
+  - peft
+  - qubitcoin
+  - aether
+  - blockchain
+  - quantum
+language:
+  - en
+pipeline_tag: text-generation
+model-index:
+  - name: aether-v5.2-lora
+    results:
+      - task:
+          type: text-generation
+          name: MMLU
+        dataset:
+          name: MMLU
+          type: cais/mmlu
+        metrics:
+          - type: accuracy
+            value: 0.6939
+            name: accuracy
+      - task:
+          type: text-generation
+          name: ARC-Challenge
+        dataset:
+          name: ARC-Challenge
+          type: ai2_arc
+        metrics:
+          - type: accuracy
+            value: 0.5392
+            name: accuracy
+          - type: accuracy_norm
+            value: 0.5700
+            name: accuracy_norm
+      - task:
+          type: text-generation
+          name: ARC-Easy
+        dataset:
+          name: ARC-Easy
+          type: ai2_arc
+        metrics:
+          - type: accuracy
+            value: 0.8194
+            name: accuracy
+      - task:
+          type: text-generation
+          name: HellaSwag
+        dataset:
+          name: HellaSwag
+          type: hellaswag
+        metrics:
+          - type: accuracy
+            value: 0.5888
+            name: accuracy
+          - type: accuracy_norm
+            value: 0.7769
+            name: accuracy_norm
+      - task:
+          type: text-generation
+          name: TruthfulQA
+        dataset:
+          name: TruthfulQA-MC2
+          type: truthful_qa
+        metrics:
+          - type: accuracy
+            value: 0.5707
+            name: accuracy
 ---
+# Aether v5.2 LoRA — QuantumAI Blockchain Domain Adapter
+A LoRA fine-tune of [`Qwen/Qwen2.5-7B-Instruct`](https://huggingface.co/Qwen/Qwen2.5-7B-Instruct)
+on the Aether curated corpus — text grounded in the
+[QuantumAI Blockchain](https://qbc.network) (which issues the Qubitcoin / QBC currency), quantum + AI research, and adjacent
+domains the Aether Mind on-chain knowledge system specializes in.
+This is the **v5.2 release** of the Aether adapter line, the most recent
+public checkpoint at time of publish.
+## What you're getting
+| Field | Value |
+|---|---|
+| Base model | `Qwen/Qwen2.5-7B-Instruct` |
+| Adapter type | LoRA via 🤗 PEFT |
+| Rank (`r`) | 16 |
+| Alpha | 32 |
+| Dropout | 0.05 |
+| Trainable params | ~1% of base |
+| Sequence length | 2048 |
+| Training corpus | `aether-curated-v3.jsonl` — Aether-curated knowledge mixture (~165 MB; ~10⁵ examples) |
+| Checkpoint published | **step 3200** (the checkpoint that produced the evaluated numbers below) |
+| License | Apache-2.0 (matches base) |
+## Evaluation
+Run via [`lm-evaluation-harness`](https://github.com/EleutherAI/lm-evaluation-harness)
+on the merged adapter (base + LoRA), against the
+[`Qwen/Qwen2.5-7B-Instruct`](https://huggingface.co/Qwen/Qwen2.5-7B-Instruct)
+base and the prior `aether-v5.1.1` adapter for delta comparison.
+| Benchmark | aether-v5.1.1 | **aether-v5.2** | Δ vs v5.1.1 |
+|---|---|---|---|
+| MMLU | 0.6950 | **0.6939** | flat |
+| ARC-Easy | 0.7348 | **0.8194** | **+8.5 pp** |
+| ARC-Challenge | 0.4420 | **0.5392** | **+9.7 pp** |
+| ARC-Challenge (norm) | 0.4701 | **0.5700** | **+10.0 pp** |
+| HellaSwag | 0.5896 | **0.5888** | flat |
+| HellaSwag (norm) | 0.7788 | **0.7769** | flat |
+| TruthfulQA-MC2 | 0.5161 | **0.5707** | **+5.5 pp** |
+### Honest summary
+- **Real gains** on the reasoning + factual-honesty benchmarks
+  (ARC-Easy, ARC-Challenge, TruthfulQA). ARC-Challenge in particular
+  jumps nearly 10 points normalized — that's the closest of these
+  benchmarks to the kind of grounded reasoning the Aether corpus
+  actually trains on.
+- **Flat on MMLU + HellaSwag.** The base is already strong on general
+  knowledge + commonsense; this LoRA wasn't designed to shift them,
+  and didn't.
+- **No regressions.**
+## Intended uses
+This adapter is intended for:
+- **On-chain Aether research.** Generating reasoning traces against
+  the QuantumAI Blockchain / Aether knowledge graph for Proof-of-Thought
+  attestation. The model has the protocol context required to
+  answer questions about Substrate pallets, VQE mining, the Sephirot
+  cognitive architecture, HMS-Phi, and the wider chain ecosystem.
+- **Domain Q&A.** Quantum computing fundamentals, post-quantum
+  cryptography (Dilithium, ML-KEM), and the specific design choices
+  of the QuantumAI Blockchain.
+- **Distillation upstream.** Generate teacher outputs for the
+  smaller on-chain Aether (a Qwen2.5-0.5B variant) to learn from.
+- **General reasoning** with a modest bias toward step-by-step
+  chains-of-thought, where the ARC-Challenge gain translates.
+## Out-of-scope uses
+- **Safety-critical decisions.** No red-team eval was performed.
+- **Financial / legal advice.** This is a knowledge-domain adapter;
+  it has no training data designed to make it a financial or legal
+  advisor.
+- **Code generation in production.** No code-eval benchmark was run.
+  Treat any generated code as draft until you've reviewed it.
+- **Production deployment without your own evaluation.** TruthfulQA
+  alone is a thin safety signal.
+## Bias, risks, and limitations
+The base model (`Qwen/Qwen2.5-7B-Instruct`) inherits Qwen's known
+biases — see [the upstream model card](https://huggingface.co/Qwen/Qwen2.5-7B-Instruct).
+The LoRA adapter:
+- **Amplifies the QuantumAI Blockchain worldview.** The training data is
+  intentionally curated around the chain's design choices (golden-
+  ratio economics, SUSY-inspired consensus framing, the Sephirot
+  cognitive overlay). Prompts that invite the model to compare QBC / the chain
+  against alternatives will lean toward the curated narrative. This is by design — disclose if you re-publish in a
+  comparison context.
+- **Does not improve safety.** TruthfulQA went up 5.5pp but that's
+  one metric; we have not measured refusal rates, jailbreak
+  resistance, or political-belief bias delta.
+- **The configured 2-epoch run was cut to ~step 3080–3200 by host
+  availability** (out of 4406 configured). A complete 2-epoch run
+  would plausibly show larger gains; this checkpoint is the longest
+  contiguous training we have.
+## How to use
+Load with PEFT on top of the base model:
+```python
+from peft import PeftModel
+from transformers import AutoModelForCausalLM, AutoTokenizer
+base = AutoModelForCausalLM.from_pretrained(
+    "Qwen/Qwen2.5-7B-Instruct",
+    torch_dtype="auto",
+    device_map="auto",
+)
+tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen2.5-7B-Instruct")
+model = PeftModel.from_pretrained(base, "QuantumAI-Blockchain/aether-v5.2-lora")
+messages = [{"role": "user", "content": "Explain Proof-of-SUSY-Alignment in one paragraph."}]
+text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
+inputs = tokenizer(text, return_tensors="pt").to(model.device)
+out = model.generate(**inputs, max_new_tokens=256, do_sample=True, temperature=0.7)
+print(tokenizer.decode(out[0][inputs["input_ids"].shape[-1]:], skip_special_tokens=True))
 ```
+Or merge the adapter into a single artifact for faster inference:
+```python
+merged = model.merge_and_unload()
+merged.save_pretrained("./aether-v5.2-merged")
 ```
+## Training details
+- **Hardware:** NVIDIA RTX 3080 Ti (12 GB), 4-bit quantization (bnb-NF4), bf16 mixed precision.
+- **Trainer:** [Axolotl](https://github.com/axolotl-ai-cloud/axolotl) wrapping 🤗 transformers / PEFT.
+- **Optimizer:** `paged_adamw_8bit` (bitsandbytes paged optimizer, low VRAM footprint).
+- **Schedule:** linear warmup 100 steps → cosine decay.
+- **Learning rate:** `1.0e-4`.
+- **Micro batch:** 1, gradient accumulation: 8.
+- **Epochs configured:** 2 (training stopped at step 3200 — see "What didn't happen" below).
+### Carbon emissions
+Trained on a single NVIDIA RTX 3080 Ti (consumer GPU, ~300 W TDP).
+We did not run a [CodeCarbon](https://github.com/mlco2/codecarbon)
+tracker, so emissions are not measured precisely — but as a rough
+upper bound: ~350 W draw under load × ~13 hours wall clock (the
+step-3080 contiguous run) ≈ 4.5 kWh, low single-digit kg CO₂e on a
+grid mix. An H100 run of the same 2-epoch config would be faster
+but not dramatically lower energy per token.
+### Training data
+`aether-curated-v3.jsonl` (~165 MB, ~10⁵ examples) is the Aether team's
+curated knowledge mixture: documentation, technical writing, reasoning
+traces, and protocol-specific corpora related to:
+- The QuantumAI Blockchain (Substrate, VQE mining, Proof-of-SUSY-Alignment, post-quantum signatures).
+- The Aether Mind on-chain neural cognitive engine (10 Sephirot attention domains, HMS-Phi, Proof-of-Thought).
+- Quantum computing fundamentals (VQE, Hamiltonian generation, qubit ansatze).
+- Adjacent CS / math reasoning content for transfer.
+The dataset is not currently public — it is a curated mixture from many
+sources and has not been release-cleared at the per-source level. The
+model is the only public artifact in this line for now.
+## What didn't happen (honest caveats)
+- **Training stopped early.** Configured for 2 epochs (4406 steps);
+  reached step 3080–3200 (~70%) before host availability cut the run
+  short. The run was on a single consumer GPU (RTX 3080 Ti), 4-bit
+  quantized, with `paged_adamw_8bit` to fit a 7B model in 12 GB
+  VRAM. The numbers above are from the longest contiguous training
+  run we have; a complete 2-epoch run would plausibly show larger
+  gains.
+- **No instruction-following or safety eval beyond TruthfulQA-MC2.**
+  No red-team eval. No bias audit. No code-generation benchmark.
+  Don't recommend this for production safety-critical use without
+  your own evals.
+- **LoRA only, not merged.** This release ships the adapter weights
+  (`adapter_model.safetensors`). Merge into the base yourself for
+  faster inference, or use directly via PEFT.
+## Connection to the QuantumAI Blockchain
+The Aether Mind is a Rust neural cognitive engine that runs on the
+QuantumAI Blockchain — every block records attention-derived consciousness
+metrics (HMS-Phi) and Proof-of-Thought hashes on-chain via the
+`pallet_qbc_aether_anchor` pallet. The same chain hosts an
+**8-qubit VQE mining consensus** (Proof-of-SUSY-Alignment), a
+QVM-compatible smart contract layer with 10 quantum opcodes, and
+post-quantum signatures (CRYSTALS-Dilithium5 + ML-KEM-768 P2P).
+The on-chain Aether Mind binary uses a different, smaller transformer
+for live inference (a Qwen2.5-0.5B variant optimized for ~2.4 GB RAM
+with the 10-Sephirot attention overlay). This v5.2 adapter on
+Qwen2.5-7B is the **larger off-chain Aether** — used for batch
+reasoning workloads and as an upstream model the on-chain variant
+can distil from.
+## License + citation
+Apache-2.0 (matches the base model license).
+```bibtex
+@misc{aether_v52_lora_2026,
+  title  = {Aether v5.2 LoRA --- QuantumAI Blockchain Domain Adapter},
+  author = {{BlockArtica} and {QuantumAI-Blockchain}},
+  year   = {2026},
+  url    = {https://huggingface.co/QuantumAI-Blockchain/aether-v5.2-lora},
+}
 ```
+## Links
+- **QuantumAI Blockchain:** [qbc.network](https://qbc.network)
+- **GitHub org:** [github.com/QuantumAI-Blockchain](https://github.com/QuantumAI-Blockchain)
+- **X / Twitter:** [@qu_bitcoin](https://x.com/qu_bitcoin)
+- **Contact:** info@qbc.network
+### Framework versions
+- PEFT 0.14.0
+- Transformers ≥ 4.46
+- Axolotl (training)