gifted_maths_gemma_4_gguf_v2

arabellako22/gifted_maths_gemma_4_gguf_v2

gifted_maths_gemma_4_gguf_v2 is a text generation model from arabellako22. Use it when you need the model to write or continue text. The card lists the license as gemma.

Verified Q4KM and Q5KM GGUF builds of mathsengine, a Gemma 4 model fine-tuned to produce clear, step-by-step mathematics solutions in LaTeX.

Downloads · 30 days

130

26% of all-time downloads

All-time downloads

502

Public

Repo size

7.1 GB

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.gguf7.1 GB · 100%

At a glance

Task: Text Generation
License: gemma
Access: Public
Created: Jun 21, 2026
Updated: Oct 13, 2026
SHA: 43d53fef

Try a prompt

Base models

arabellako22/gifted_maths_unsloth_remerged_full_gemma_4

Model card

From the Hugging Face model README

Gifted Maths Gemma 4 GGUF

Verified Q4_K_M and Q5_K_M GGUF builds of maths_engine, a Gemma 4 model fine-tuned to produce clear, step-by-step mathematics solutions in LaTeX.

Key facts

Field	Value
Developer	arabellako22
Original base	unsloth/gemma-4-E2B-it
LoRA adapter	arabellako22/gifted_maths_lora_model
Conversion source	arabellako22/gifted_maths_unsloth_remerged_full_gemma_4
Dataset	arabellako22/gifted_maths_training_dataset
Format	GGUF for llama.cpp-compatible runtimes
Task	Step-by-step mathematics reasoning
Language	English
License	Gemma license

The earlier Gifted-Maths-Gemma-4_legacy export does not contain the verification record used for this release and is retained only as a legacy artifact.

Files

File	Quantization	Size	Guidance
`gifted-maths-gemma-4-v2.Q4_K_M.gguf`	Q4_K_M	3.43 GB	Recommended local build
`gifted-maths-gemma-4-v2.Q5_K_M.gguf`	Q5_K_M	3.63 GB	Larger, generally higher fidelity
`Modelfile`	Ollama template	686 B	Optional Ollama helper

Prompt format

<start_of_turn>system
You are maths_engine, a local gifted mathematics reasoning model for students.
You solve maths problems step by step, clearly and calmly.
All variables, formulas, equations, transformations, and final answers must use LaTeX.
Use inline math with $...$ and display math with $$...$$.
Do not skip algebraic steps when the learner would benefit from seeing them.
End every solution with a clear line starting with: Final Answer:<end_of_turn>
<start_of_turn>user
{{ question }}<end_of_turn>
<start_of_turn>model

Run with llama.cpp

Save the formatted prompt as prompt.txt, then run:

llama-cli \
  -m gifted-maths-gemma-4-v2.Q4_K_M.gguf \
  -f prompt.txt \
  -n 900 \
  --temp 0

Deterministic decoding (--temp 0) reproduces the recorded verification path.

Verification

Stage	Recorded result
Base + adapter	Generated a worked algebra solution with a final answer
Unsloth 16-bit merge	Returned `2160` for the coefficient of $x^4$ in $(2x-3)^6$
Local Q4_K_M	Returned $x=\frac{1}{2}$ or $x=-4$ for $2x^2+7x-4=0$
Re-downloaded Q4_K_M	Returned the same correct roots after Hub download

Conversion used llama.cpp's convert_hf_to_gguf.py with an F16 intermediate, followed by llama-quantize for Q4_K_M and Q5_K_M. These are narrow artifact-integrity checks, not a mathematics benchmark.

Intended use

Local mathematics-assistant experiments
Worked-example and LaTeX-formatting demonstrations
Testing the verified Gifted Maths merge in llama.cpp-compatible applications

Limitations

The source adapter was trained on only 54 examples and validated on 7; no independent test set was created.
The provenance and licensing of the original 50 dataset examples were not recorded.
No standardized mathematics, factuality, safety, bias, or robustness benchmark is published.
Generated algebra, arithmetic, proofs, notation, or LaTeX may be wrong.
Quantization may alter output quality relative to the merged model.
“Gifted” describes the project goal, not a validated learner classification.

Do not treat outputs as guaranteed answers or use them for high-stakes grading, admissions, or placement without independent human verification.

Task: Text Generation
License: gemma
Languages: en
Created: Jun 21, 2026
Updated: Oct 13, 2026