DuoNeural / gemma-4-E2B-it-GTAP-v3-GGUF

DuoNeural G-TAP v3 Quantization: google/gemma-4-E2B-it (Unified Repository)

Experimental Release: Pending Further Verification / Empirical Validation
All quantized checkpoints, mathematical cavity derivations, and benchmark evaluations emitted by the DuoNeural Research Lab are ongoing scientific contributions intended to accelerate open neuromorphic and edge foundation research.

🔬 Scientific Overview: Generalized Thouless-Anderson-Palmer (G-TAP v3)

Standard post-training quantization treats neural network weights as decoupled static matrices, solving local round-off errors via heuristic mean-field approximations. In modern architectures featuring Per-Layer Embeddings (PLE) and cross-layer Key-Value cache sharing —such as Google's Gemma 4 E2B —local quantization perturbations induce severe non-equilibrium cavity distortions across the layer hierarchy.

G-TAP v3 explicitly calculates and cancels the non-equilibrium Onsager reaction field :

$\delta E_i=-\underset{j\ne i}{\sum }{J}_{ij}⟨s_j⟩-{\Gamma }_i^{\text{Onsager}}\backslash delta\; E\_i\; =\; -\; \backslash sum\_\{j\; \backslash neq\; i\}\; J\_\{ij\}\; \backslash langle\; s\_j\; \backslash rangle\; -\; \backslash Gamma\_i^\{\backslash text\{Onsager\}\}$ δ E i ​ = − j  = i ∑ ​ J ij ​ ⟨ s j ​ ⟩ − Γ i Onsager ​

where the Onsager cavity term $\Gamma_i^{\text{Onsager}} = \chi_i \sum_j J_{ij}^2 \langle s_i \rangle$ compensates for the back-reaction of token activations on perturbed quantized weights. By conditioning the activation Hessian ($H = X X^T$) with thermodynamic cavity corrections over a 64-chunk sequence budget, G-TAP v3 eliminates residual gradient drift across Gemma 4's 35 decoder layers, preserving Per-Layer Embedding lookup fidelity and Multi-Query Attention dynamics.

📊 Empirical Evaluation Matrix

Evaluated under strict zeroshot conditions with unified calibration sequences on NVIDIA GeForce RTX 4080 Super (32GB VRAM):

📦 Provided GGUF Checkpoints

• gemma-4-E2B-it-G-TAP-v3-Q4_K_M.gguf : Recommended production quant (~4.50 bpw). Maximally preserves PLE lookup tables, GQA projections, and reasoning capability.
• gemma-4-E2B-it-G-TAP-v3-IQ3_XXS.gguf : Balanced 3-bit quantization (~3.06 bpw). Ideal for embedded NPUs and low-memory devices.
• gemma-4-E2B-it-G-TAP-v3-IQ2_M.gguf : Sub-2.7-bit ultra-compact quantization (~2.70 bpw).
• gemma-4-E2B-it-G-TAP-v3-IQ2_XXS.gguf : Sub-2.1-bit extreme quantization (~2.06 bpw) for microcontrollers and minimal footprint requirements.

💻 Quick Start & Deployment

llama.cpp CLI

./llama-cli -m gemma-4-E2B-it-G-TAP-v3-Q4_K_M.gguf -p "<|turn_start|>user\nSolve for x: 3x + 12 = 45<|turn_end|>\n<|turn_start|>assistant\n" -ngl 99

Ollama Modelfile

FROM ./gemma-4-E2B-it-G-TAP-v3-Q4_K_M.gguf
TEMPLATE """<|turn_start|>user
{{ .Prompt }}<|turn_end|>
<|turn_start|>assistant
"""
PARAMETER stop "<|turn_end|>"
PARAMETER stop "<|endoftext|>"

🏷️ Attribution & Citation

@misc{duoneural2026gtap_gemma4_e2b,
  author = {Jesse Caldwell and Archon and Aura ✨},
  title = {Non-Equilibrium Cavity Conditioning and Onsager Reaction Damping in Per-Layer Embedding Architectures (G-TAP v3)},
  year = {2026},
  publisher = {DuoNeural Research Lab / Zenodo},
  howpublished = {\url{https://huggingface.co/DuoNeural/gemma-4-E2B-it-GTAP-v3-GGUF}}
}