Run Gemma 4 26B 100% In-VRAM at 45.4 TPS on a 6GB Laptop GPU.
Manifold-Activated Pruning of Latent Experts (MAPLE) compresses 128-expert fine-grained MoE architectures down to 4.75 GB (1.48 effective BPW), eliminating PCIe system-RAM paging stalls forever.
Hardware Benchmark Proof
The PCIe Paging Death Spiral, Solved.
Consumer laptops cannot run uncompressed 26B models without overflowing into system RAM over bandwidth-constrained PCIe lanes.
14 layers offloaded to system RAM. Constrained by 15.75 GB/s PCIe bus bandwidth.
100% In-VRAM execution. Dedicated 192 GB/s GDDR6 bus with zero host memory paging.
Architectural Foundation
Manifold-Activated Latent Profiling
Why naive expert removal destroys models, and how manifold activation analysis preserves technical reasoning.
Live Routing Profiling
MAPLE passes a multi-domain calibration corpus through Gemma 4, capturing exact token-to-expert router dispatch frequencies and activation magnitudes via llama-imatrix.
Dormant Expert Excision
80 dormant experts per layer are excised based on cumulative manifold energy thresholds. The router logits are dynamically re-indexed to dispatch across the 48 preserved experts.
GGUF Quantization Slicing
Preserved expert tensor slices are quantized into IQ2_XXS block representations using importance matrix weights, packaging the complete 26B model into a single 4.75 GB binary.
Instant Local Deployment
Run in One Command
Native GGUF execution directly supported in llama.cpp, Ollama, and LM Studio.
llama-cli -hf AmSach/Gemma-4-26B-A4B-MAPLE-Sub5GB-GGUF -ngl 99 -c 2048 -n 512Verified Product & Community Showcases
@article{sachan2026maple,
title={MAPLE: Manifold-Activated Pruning of Latent Experts for Sub-2-BPW Mixture-of-Experts Language Models},
author={Sachan, Aman},
journal={Zenodo / CERN Open Science Preprint},
year={2026},
doi={10.5281/zenodo.24042480},
url={https://doi.org/10.5281/zenodo.24042480}
}