magnitudedev/Muse-Glimmer-30B-DFlash-GGUF overview
Muse Glimmer 30B DFlash GGUF Metadata corrected copy of dflash kquant.gguf from unsloth/Muse Glimmer 30B GGUF https://huggingface.co/unsloth/Muse Glimmer 30B G…
Runs locally from ~1.52 GB disk (4 GB VRAM class GPUs with llama.cpp / guIDE).
Repository Files & Downloads
| File | Type | Quantization | Size | Link |
|---|---|---|---|---|
| dflash-kquant.gguf | GGUF | GGUF | 1.52 GB | Download |
Model Details
| Model ID | magnitudedev/Muse-Glimmer-30B-DFlash-GGUF |
|---|---|
| Author | magnitudedev |
| Pipeline | — |
| License | apache-2.0 |
| Base model | meta-models/Muse-Glimmer-30B-Assistant |
| Last modified | 2026-10-06T22:07:02.000Z |
Model README
---
license: apache-2.0
base_model: meta-models/Muse-Glimmer-30B-Assistant
tags:
- gguf
- dflash
- speculative-decoding
- magnitude
inference: false
---
Muse Glimmer 30B DFlash GGUF
Metadata-corrected copy of dflash-kquant.gguf from unsloth/Muse-Glimmer-30B-GGUF, pinned at 1afeb8e879f60116d206cf724425dbe1e1a2f7f5. The underlying assistant is Meta's Muse Glimmer 30B Assistant; this model requires its matching target model.
SHA-256: 66d69be5fcd16d60b00c87a9b92bb77c47d325291df891363c29c96428a9dc5c
Draft metadata
dflash.attention.causal = false: the draft block uses bidirectional attention.dflash.sample_from_anchor = false: proposals come from mask rows, starting after the anchor.dflash.attention.sliding_window = 2049: strict-distance bound preserving the upstream inclusive radius of 2048.- Derived from the checkpoint configuration and Transformers reference implementation; verified against reference outputs.
Run magnitudedev/Muse-Glimmer-30B-DFlash-GGUF with guIDE
Download guIDE — the AI-native code editor with local LLM inference and 69 built-in tools.
Source: Hugging Face · Compare models