gemma-4-E4B-RotorQuant-GGUF-Q2_K

GGUF Q2_K weight-quantized variant of google/gemma-4-E4B optimised for use with RotorQuant KV cache compression via a dedicated llama.cpp fork.

Important: RotorQuant KV cache types (planar3, iso3) are not available in upstream llama.cpp, standard Ollama, or LM Studio. They require a specific llama.cpp fork. The GGUF file itself is a standard GGUF and works with any llama.cpp-compatible runtime using normal KV cache types (f16, q8_0, q4_0, etc.).

Hardware compatibility

Device	VRAM / RAM	Recommendation
CPU host with ≥8 GB RAM	~1.5 GB	works via llama.cpp; slower than GPU but no accelerator required
Apple Silicon (Metal)	~1.7 GB	llama.cpp Metal backend; fast on M-series unified memory
NVIDIA GPU (partial offload)	split between GPU + RAM	offload as many layers as VRAM allows; rest on CPU

Overview

This model combines two independent compression techniques:

Technique	What it does	Requirement
GGUF Q2_K weight quantization	Reduces model size from ~8 GB (BF16) to ~1.4 GB	Any llama.cpp-compatible runtime
RotorQuant KV cache compression — block-diagonal Clifford-algebra rotors for 3-bit KV cache (--cache-type-k iso3 --cache-type-v iso3)	Block-diagonal rotations / random rotation for compressed KV cache	llama-cpp-turboquant fork only

Quickstart

Option A — With RotorQuant KV cache (fork required)

You must build from the RotorQuant-enabled llama.cpp fork:

# Clone and build the fork
git clone https://github.com/johndpope/llama-cpp-turboquant.git
cd llama-cpp-turboquant && git checkout feature/planarquant-kv-cache

# CUDA (Windows/Linux)
cmake -B build -DGGML_CUDA=ON -DCMAKE_BUILD_TYPE=Release && cmake --build build -j

# Metal (Apple Silicon)
cmake -B build -DGGML_METAL=ON -DGGML_METAL_EMBED_LIBRARY=ON -DCMAKE_BUILD_TYPE=Release && cmake --build build -j

# Run with RotorQuant KV cache
./build/bin/llama-cli -m gemma-4-E4B-RotorQuant-GGUF-Q2_K.gguf \
  --cache-type-k iso3 --cache-type-v iso3 \
  -ngl 99 -fa \
  -p "Explain quantum computing"

# Or run as a server
./build/bin/llama-server -m gemma-4-E4B-RotorQuant-GGUF-Q2_K.gguf \
  --cache-type-k iso3 --cache-type-v iso3 \
  -ngl 99 -fa --jinja

Option B — With standard llama.cpp / LM Studio / Ollama

The GGUF works as a normal quantised model. You won't get RotorQuant-specific KV cache benefits, but standard KV cache quantization (q8_0, q4_0) still reduces VRAM significantly.

llama.cpp (upstream)

llama-cli -m gemma-4-E4B-RotorQuant-GGUF-Q2_K.gguf \
  --cache-type-k q8_0 --cache-type-v q8_0 \
  -ngl 99 -fa \
  -p "Explain quantum computing"

LM Studio

1. Download the GGUF file and load in LM Studio.
2. Enable Developer Mode (Settings → Developer).
3. In the model loader's advanced settings, set Flash Attention to ON.
4. Set K Cache Quantization and V Cache Quantization to q8_0 (or q4_0 for more aggressive VRAM savings).
5. Note: LM Studio does not currently support RotorQuant's iso3 cache types. Track this feature request for updates.

Ollama

# Standard Ollama does not support RotorQuant cache types.
# Use with default or q8_0 KV cache via OLLAMA_KV_CACHE_TYPE=q8_0
OLLAMA_KV_CACHE_TYPE=q8_0 OLLAMA_FLASH_ATTENTION=1 ollama run majentik/gemma-4-E4B-RotorQuant-GGUF-Q2_K

Specifications

Property	Value
Base Model	google/gemma-4-E4B
Architecture	Dense transformer (Edge optimised)
Parameters	~4B
Context Length	128K
Weight Quantization	GGUF Q2_K (aggressive 2-bit, noticeable quality drop)
Original Size (BF16)	~8 GB
Quantized File Size	~1.4 GB
KV Cache (RotorQuant)	3-bit via --cache-type-k iso3 --cache-type-v iso3 (fork only)
KV Cache (standard)	q8_0, q4_0, f16, etc. (any llama.cpp runtime)
License	apache-2.0
Modalities	Text + Image + Audio
Compatible Runtimes	llama.cpp, LM Studio, Ollama, koboldcpp

What is RotorQuant?

RotorQuant is a KV cache compression method based on Clifford algebra (Cl(3,0)) rotors. It was developed as a faster, more parameter-efficient alternative to Google's TurboQuant (ICLR 2026).

Instead of applying a dense d×d random orthogonal rotation matrix (as TurboQuant does), RotorQuant uses lightweight block-diagonal rotations — independent 2D/4D rotations per pair/quartet — achieving O(d) complexity instead of O(d log d), fully parallelisable with no inter-element dependencies.

Benchmarks from the RotorQuant repository (Llama 3.1 8B, RTX 5090 — results will vary by model and hardware):

Metric	RotorQuant (iso3)	TurboQuant	Standard q4_0
Prefill Speed	3,822 tok/s	722 tok/s	—
Decode Speed	119 tok/s	93 tok/s	—
Perplexity (PPL)	6.91	7.07	—
KV Compression	~5× vs FP16	~5× vs FP16	~4× vs FP16
Rotation Parameters	4 per rotor	16,384 per matrix	N/A

Note: These benchmarks are from the RotorQuant repository using Llama 3.1 8B on an RTX 5090. Performance on gemma-4-E4B will differ. Independent benchmarks for this specific model are welcome — please open a discussion if you have results to share.

Current Status of RotorQuant in the Ecosystem

Runtime	RotorQuant Support	Standard KV Quant
llama.cpp (upstream)	❌ Not merged	✅ q8_0, q4_0, q4_1, iq4_nl, q5_0, q5_1
llama-cpp-turboquant fork	✅ planar3, iso3	✅ All standard types
LM Studio	❌ Requested	✅ Via advanced settings
Ollama	❌ Not supported	✅ Via OLLAMA_KV_CACHE_TYPE
koboldcpp	❌ Not supported	✅ Standard types

Recommended Settings

For VRAM-constrained setups, standard q8_0 KV cache quantization already halves KV cache memory with negligible quality impact. Flash Attention should always be enabled — it is required for V cache quantization and improves memory efficiency regardless.

VRAM	Suggested Configuration
24 GB (RTX 4090)	Q2_K + q8_0 KV cache + Flash Attention, 8K–16K context
16 GB	Q2_K + q4_0 KV cache + Flash Attention, 4K–8K context
48+ GB	Q2_K + f16 KV cache, full 32K+ context

See Also

• RotorQuant GitHub
• llama-cpp-turboquant fork
• TurboQuant llama.cpp discussion
• TurboQuant paper (arXiv: 2504.19874)
• Base model: google/gemma-4-E4B
• gemma-4-E4B announcement

Quant trade-off (GGUF lane)

Quant	Approx size	Use case	Recommendation
Q2_K	~2.2 GB	Lossy, low-RAM CPU/edge	Resource-constrained inference
Q3_K_M	~2.4 GB	Smaller-than-Q4, modest quality drop	Edge devices with ~16 GB RAM
IQ4_XS	~2.1 GB	Importance-quant 4-bit, smaller than Q4_K_M	Best size/quality at 4-bit
Q4_K_M	~3.0 GB	Balanced default	Recommended for most users
Q5_K_M	~3.1 GB	Higher fidelity than Q4	Quality-sensitive applications
Q6_K	~3.6 GB	Approaching FP16 quality	High-fidelity CPU/edge
Q8_0	~4.1 GB	Near-lossless reference	Fidelity-critical work
MXFP4_MOE	~2.2 GB	Microscaling FP4 (MoE-aware)	vLLM / transformers users

(Current variant — Q2_K — is bolded.)

Variants in this family

(Showing 18 sibling variants under majentik/gemma4-e4b-*. The current variant — RotorQuant-GGUF-Q2_K — is bolded.)

Variant	Runtime	Approx size	Use case
RotorQuant	runtime modifier	n/a	KV-cache root (weight-agnostic)
RotorQuant-AWQ-4bit	transformers	~2.5 GB	GPU 4-bit (AutoAWQ)
RotorQuant-AWQ-8bit	transformers	~4.4 GB	GPU 8-bit (AutoAWQ)
RotorQuant-GGUF-IQ4_XS	llama.cpp	~3.4 GB	Lossy 4-bit, low-RAM CPU/edge
RotorQuant-GGUF-Q2_K	llama.cpp	~2.4 GB	Lossy, low-RAM CPU/edge
RotorQuant-GGUF-Q3_K_M	llama.cpp	~3.1 GB	Smaller 3-bit, CPU-friendly
RotorQuant-GGUF-Q4_K_M	llama.cpp	~4.4 GB	Balanced default
RotorQuant-GGUF-Q5_K_M	llama.cpp	~5.3 GB	Higher fidelity, more RAM
RotorQuant-GGUF-Q8_0	llama.cpp	~8.4 GB	Near-lossless reference
RotorQuant-MLX-2bit	mlx-lm	~1.3 GB	Apple Silicon, smallest
RotorQuant-MLX-4bit	mlx-lm	~2.5 GB	Apple Silicon balanced
RotorQuant-MLX-8bit	mlx-lm	~4.7 GB	Apple Silicon reference
TurboQuant	runtime modifier	n/a	KV-cache root (weight-agnostic)
TurboQuant-AWQ-4bit	transformers	~2.5 GB	GPU 4-bit (AutoAWQ)
TurboQuant-AWQ-8bit	transformers	~4.4 GB	GPU 8-bit (AutoAWQ)
TurboQuant-MLX-2bit	mlx-lm	~1.3 GB	Apple Silicon, smallest
TurboQuant-MLX-4bit	mlx-lm	~2.5 GB	Apple Silicon balanced
TurboQuant-MLX-8bit	mlx-lm	~4.7 GB	Apple Silicon reference