GraySoft
Projects Models Compare Cloud benchmarks FAQ Download guIDE →
Model Intelligence Sheet

TaQuants/Gemma-4-12B-OBLITERATED-TaQuants-GGUF overview

Gemma 4 12B OBLITERATED TaQuants The Repository https://github.com/ek15072809/TaQuants Technical Report https://github.com/ek15072809/TaQuants/blob/main/docs/T…

ggufTaQuantsuncensorednon-censoredunfilteredtext-generationbase_model:OBLITERATUS/Gemma-4-12B-OBLITERATEDbase_model:quantized:OBLITERATUS/Gemma-4-12B-OBLITERATEDendpoints_compatibleregion:usimatrixconversational

Runs locally from ~5.48 GB disk (8 GB VRAM class GPUs with llama.cpp / guIDE).

Downloads
332
Likes
0
Pipeline
text-generation
Author

Repository Files & Downloads

1 GGUF files detected
Direct downloads for local inference
FileTypeQuantizationSizeLink
Gemma-4-12B-OBLITERATED-TaIQ3_S.ggufGGUFGGUF5.48 GBDownload

Model Details

Model IDTaQuants/Gemma-4-12B-OBLITERATED-TaQuants-GGUF
AuthorTaQuants
Pipelinetext-generation
License
Base modelOBLITERATUS/Gemma-4-12B-OBLITERATED
Last modified2026-07-03T10:49:59.000Z

Model README

---

base_model:

  • OBLITERATUS/Gemma-4-12B-OBLITERATED

pipeline_tag: text-generation

tags:

  • TaQuants
  • uncensored
  • non-censored
  • unfiltered

---

Gemma-4-12B-OBLITERATED TaQuants

The Repository

Technical Report

The Tema_Q development team, team zenei, has developed a new importance matrix method called TaQuants (Tensor-aware Adaptive Quantization).

This model is a TaQuants version of OBLITERATUS/Gemma-4-12B-OBLITERATED created with TaQuants v3.0.

The model size and performance are as follows:

| | Size (GB) | Knowledge score |

| --- | --- | --- |

| Q4_K_M | 7.38 | 31.97 |

| TaIQ3_S | 5.48 | 31.88 |

| TaIQ2_M | | 9.09 |

Vulnerabilities to quantization were identified in the model. Q4_K_M caused partial collapse, while TaIQ2_M exhibited complete collapse. TaIQ3_S performs inference comparable to Q4_K_M but avoids the partial collapse observed with Q4_K_M.

Run TaQuants/Gemma-4-12B-OBLITERATED-TaQuants-GGUF with guIDE

Download guIDE — the AI-native code editor with local LLM inference and 69 built-in tools.

Download guIDE → · Browse 524k+ models · Compare models

Source: Hugging Face · Compare models