gemma-4-E2B-it-qat-GGUF

JMingo/gemma-4-E2B-it-qat-GGUF

gemma-4-E2B-it-qat-GGUF is a machine learning model from JMingo. Use it for the machine learning task on the model card, and read the license before you ship it in a product. The card lists the license as apache-2.0.

Downloads · 30 days

155

3% of all-time downloads

All-time downloads

4.8K

Public

Repo size

25.6 GB

Likes

Public

Hugging Face

Repo makeup

Click a slice to open those files.

.gguf13.5 GB · 100%

At a glance

License: apache-2.0
Access: Public
Created: Jun 12, 2026
Updated: Oct 1, 2026
SHA: ccff9a1c

Base models

google/gemma-4-E2B-it-qat-q4_0-unquantized

Model card

From the Hugging Face model README

Near-lossless GGUF quants of:

Chat template has been updated (2026-07-15 version)

Original QAT Specifications

The base QAT model was trained with the following specifications:

Text & Draft Models: 4-bit
Multimodal Projector (mmproj):
- Audio Encoder: Mostly 2-bit (with some tensors in 4-bit or BF16)
- Vision Encoder: 8-bit
- Other parts: BF16

Quantization Error Evaluation

Dequantization: Quantized weights were dequantized to F32 for evaluation.
Error Calculation: Quantization error metrics (compared to the BF16 baseline) were computed and accumulated in F64 precision to prevent numerical underflow and precision loss.

Metrics:

MAE (Mean Absolute Error)
RMSE (Root Mean Squared Error)
Max Error

[!NOTE] Interpretation of Metrics: These error metrics do not directly reflect actual model performance. This is because imatrix optimization intentionally increases measured error by downweighting less important parameters to improve overall performance. However, these metrics remain useful for estimating how close the quantized model is to the original, lossless quality.

gemma-4-E2B-it-qat

Model	MAE	RMSE	Max Error
`JMingo/gemma-4-E2B-it-qat-BF16.gguf` (Baseline)	0.00000000	0.00000000	0.00000000
`JMingo/gemma-4-E2B-it-qat-Q4_0.gguf`	0.00003548	0.00006434	0.00256348
`idkwhattoputherenow/gemma-4-E2B-it-qat-q4_0-unquantized-q4_0-maxerr.gguf`	0.00004027	0.00007371	0.00299072
`unsloth/gemma-4-E2B-it-qat-UD-Q4_K_XL.gguf`	0.00005024	0.00008256	0.00341797
`google/gemma-4-E2B_q4_0-it.gguf`	0.00052373	0.00093318	0.06982422
`lmstudio-community/gemma-4-E2B-it-QAT-Q4_0.gguf`	0.00052373	0.00093318	0.06982422
`mradermacher/gemma-4-E2B-it-qat-q4_0-unquantized.i1-Q4_0.gguf`	0.00032190	0.00060744	0.03930664
`mradermacher/gemma-4-E2B-it-qat-q4_0-unquantized.Q4_K_S.gguf`	0.00054283	0.00086672	0.03857422
`mradermacher/gemma-4-E2B-it-qat-q4_0-unquantized.Q4_K_M.gguf`	0.00053254	0.00086113	0.03857422
`mradermacher/gemma-4-E2B-it-qat-q4_0-unquantized.i1-Q4_1.gguf`	0.00048704	0.00084594	0.13549805
`mradermacher/gemma-4-E2B-it-qat-q4_0-unquantized.i1-Q4_K_S.gguf`	0.00053921	0.00086299	0.14585495
`mradermacher/gemma-4-E2B-it-qat-q4_0-unquantized.i1-Q4_K_M.gguf`	0.00052917	0.00085745	0.14585495
`mradermacher/gemma-4-E2B-it-qat-q4_0-unquantized.IQ4_XS.gguf`	0.00078202	0.00116219	0.05517125
`mradermacher/gemma-4-E2B-it-qat-q4_0-unquantized.i1-IQ4_XS.gguf`	0.00075868	0.00114195	0.31855154
`mradermacher/gemma-4-E2B-it-qat-q4_0-unquantized.i1-IQ4_NL.gguf`	0.00075354	0.00113171	0.35578442
`mradermacher/gemma-4-E2B-it-qat-q4_0-unquantized.Q3_K_L.gguf`	0.00113207	0.00216116	0.12817383
`mradermacher/gemma-4-E2B-it-qat-q4_0-unquantized.i1-Q3_K_L.gguf`	0.00109405	0.00208034	0.31637573

mtp-gemma-4-E2B-it-qat

Model	MAE	RMSE	Max Error
`JMingo/mtp-gemma-4-E2B-it-qat-BF16.gguf` (Baseline)	0.00000000	0.00000000	0.00000000
`JMingo/mtp-gemma-4-E2B-it-qat-Q4_0.gguf`	0.00005037	0.00008262	0.00219727
`unsloth/mtp-gemma-4-E2B-it.gguf`	0.00007164	0.00010655	0.00781250

mmproj-gemma-4-E2B-it-qat

Model	MAE	RMSE	Max Error
`JMingo/mmproj-gemma-4-E2B-it-qat-BF16.gguf` (Baseline)	0.00000000	0.00000000	0.00000000
`JMingo/mmproj-gemma-4-E2B-it-qat-Q2_K_XL.gguf`	0.00001210	0.00003352	0.00271225

License: apache-2.0
Created: Jun 12, 2026
Updated: Oct 1, 2026