---

base_model:

- RedHatAI/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B.dspark

license: other

license_name: openmdw-1.1

license_link: https://openmdw.ai/license/1-1/

library_name: llama.cpp

tags:

- gguf

- dspark

- speculative-decoding

- magnitude

inference: false

---

NVIDIA Nemotron 3.5 Super VL 120B A12B DSpark GGUF

Q8_0 GGUF conversion of

RedHatAI/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B.dspark

for use with Magnitude.

Converted with

llama.cpp

at revision 620d2225cc436ab0b001e3200e590e3b286ef9ab.

Artifact

• File: NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-DSpark-Q8_0.gguf
• Quantization: Q8_0
• SHA-256: da49fd97d494d775285b5df78b27f673c96b4138dab0f7b5859a8b37d298066c
• Source revision: 4ff45128272e4e5c00ab82d2886946650229ced7
• Target tokenizer revision: a1ccf94d69377bf82c4db39798aac28447d9c16f

Draft metadata

• dflash.attention.causal = true: the draft block uses causal attention.
• dflash.sample_from_anchor = true: proposal selection starts at the anchor row.
• dflash.attention.sliding_window = 2048 on every layer.
• Derived from the Red Hat checkpoint configuration.