fraQtl/Gemma-4-E4B-it-Hi-Fi-GGUF overview
Gemma 4 E4B it Q4 K M GGUF — fraQtl Hi Fi calibration aware quantization Calibration aware Q4 K M build of google/gemma 4 E4B it . A modest but real fidelity i…
Runs locally from ~4.27 GB disk (8 GB VRAM class GPUs with llama.cpp / guIDE).
Repository Files & Downloads
Model Details
| Model ID | fraQtl/Gemma-4-E4B-it-Hi-Fi-GGUF |
|---|---|
| Author | fraQtl |
| Pipeline | text-generation |
| License | gemma |
| Base model | google/gemma-4-E4B-it |
| Last modified | 2026-09-16T19:00:38.000Z |
Model README
---
license: gemma
base_model: google/gemma-4-E4B-it
pipeline_tag: text-generation
tags:
- gguf
- quantized
- calibration-aware
- gemma
- on-device
---
Gemma-4-E4B-it Q4_K_M GGUF — fraQtl Hi-Fi (calibration-aware quantization)
Calibration-aware Q4_K_M build of google/gemma-4-E4B-it. A modest but real
fidelity improvement over the community Q4_K_M, strongest on code/math. Every
number below is measured, 3 independent runs (run-to-run drift 0.0), and
reproducible from the receipts in this repo.
Fidelity to the ORIGINAL bf16 model (KLD, lower = closer)
| Arm | bytes | KLD code/math | KLD general | top-1 agree (c/m) | top-1 agree (gen) |
|---|---|---|---|---|---|
| fraQtl Hi-Fi (iso-size) | 5.40 GB | 0.0387 (−16.2%) | 0.0593 (−4.2%) | 0.948 | 0.910 |
| bartowski Q4_K_M (baseline) | 5.40 GB | 0.0462 | 0.0619 | 0.948 | 0.910 |
Honest read — stated plainly, not buried: the win is real on **code/math
(−16.2%), near-parity on general (−4.2%), and top-1 agreement is a tie**
(0.948 / 0.910 both). This is a smaller lift than our Gemma-4 E2B build
(−36% / −27%) — E4B's larger footprint leaves less quantization headroom. If
your workload is code/math-heavy on-device, this helps; on general chat it's
about even with the community build.
Also included
- Phone build (
...-Phone.gguf): ~4.59 GB (~15% smaller), PLE demoted to Q4_K,
speed-neutral by construction (bandwidth-relevant bytes identical).
- GSM8K (0-shot instructed chain-of-thought, n=200, seed 0): comparable to the
community build (bartowski 84.0%). No win claim.
- Wikitext-2 perplexity: sanity flag only, not a gate.
Reproducibility
Gates 1/2 PASS + phone smoke PASS. Fidelity metric: symmetric top-20 KL divergence
on a Q8 teacher of the ORIGINAL bf16 checkpoint (the axis a user migrating from
bf16 experiences). The imatrix, recipe hash, and every artifact sha256 ship in
this repo. Runs via llama.cpp / the GenieX llama.cpp plugin (on-device / edge:
Snapdragon X-class laptops, phones).
Built by fraQtl. Community builders credited: bartowski Q4_K_M baseline.
Run fraQtl/Gemma-4-E4B-it-Hi-Fi-GGUF with guIDE
Download guIDE — the AI-native code editor with local LLM inference and 69 built-in tools.
Source: Hugging Face · Compare models