0ppxnhximxr/Muse-Glimmer-30B-Heretic-UD-Q5_K_XL-GGUF overview
Muse Glimmer 30B Heretic UD Q5 K XL GGUF This repository contains one GGUF quantization of Muse Glimmer 30B Heretic. It was quantized directly from the Heretic…
Runs locally from ~20.29 GB disk (24 GB VRAM class GPUs with llama.cpp / guIDE).
Repository Files & Downloads
| File | Type | Quantization | Size | Link |
|---|---|---|---|---|
| Muse-Glimmer-30B-Heretic-UD-Q5_K_XL.gguf | GGUF | Q5_K_XL | 20.29 GB | Download |
Model Details
| Model ID | 0ppxnhximxr/Muse-Glimmer-30B-Heretic-UD-Q5_K_XL-GGUF |
|---|---|
| Author | 0ppxnhximxr |
| Pipeline | — |
| License | apache-2.0 |
| Base model | darkc0de/Muse-Glimmer-30B-heretic |
| Last modified | 2026-09-15T12:25:58.000Z |
Model README
---
license: apache-2.0
base_model: darkc0de/Muse-Glimmer-30B-heretic
tags:
- gguf
- llama.cpp
- quantization
- muse-glimmer
---
Muse-Glimmer-30B-Heretic UD-Q5_K_XL GGUF
This repository contains one GGUF quantization of Muse-Glimmer-30B-Heretic.
It was quantized directly from the Heretic BF16 GGUF using a new importance matrix calculated on that BF16 model. No Q8 intermediate was used.
The tensor types match Unsloth's Muse-Glimmer UD-Q5_K_XL tensor by tensor: 104 Q5_K, 150 Q6_K, 164 Q8_0, and 313 F32 tensors. Calibration used 166 chunks of 512 tokens from the bartowski imatrix corpus.
| File | Size | SHA256 |
|---|---:|---|
| Muse-Glimmer-30B-Heretic-UD-Q5_K_XL.gguf | 21.79 GB | 74e3fb3995cf9ce906ff809c55f426b97dae2cd46c87c4b04100813adb73cb0b |
hf download 0ppxnhximxr/Muse-Glimmer-30B-Heretic-UD-Q5_K_XL-GGUF Muse-Glimmer-30B-Heretic-UD-Q5_K_XL.gguf
llama-cli -m Muse-Glimmer-30B-Heretic-UD-Q5_K_XL.gguf -ngl 99The Heretic model was created by darkc0de from Meta's Muse-Glimmer-30B. This release follows the upstream Apache 2.0 license.
Run 0ppxnhximxr/Muse-Glimmer-30B-Heretic-UD-Q5_K_XL-GGUF with guIDE
Download guIDE — the AI-native code editor with local LLM inference and 69 built-in tools.
Source: Hugging Face · Compare models