majentik's picture
feat: upload TurboQuant-MLX-8bit-TQ-KV combo card
cdaa881 verified
|
raw
history blame
2.24 kB
metadata
license: other
license_name: nvidia-open-model-license
license_link: >-
  https://developer.download.nvidia.com/licenses/nvidia-open-model-license-agreement-june-2024.pdf
base_model: nvidia/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-BF16
tags:
  - nemotron
  - multimodal
  - mamba2
  - moe
  - quantized
  - turboquant
  - mlx
  - kv-cache-modifier

Nemotron-3-Nano-Omni-30B-A3B-Reasoning - TurboQuant MLX 8-bit + TurboQuant KV-Cache (matched stack)

Documentation card for the matched TurboQuant weight + TurboQuant KV-cache stack of Nemotron-3-Nano-Omni-30B-A3B-Reasoning at MLX 8-bit.

No new weights are published here. Load the weights from majentik/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-TurboQuant-MLX-8bit and apply the TurboQuant KV-cache modifier documented in majentik/Nemotron-3-Nano-Omni-30B-A3B-Reasoning-TurboQuant.

Modality matrix

Modality Encoder Quantization in this variant
Text LLM backbone (Mamba-2 + Transformer hybrid Sparse MoE) per the variant suffix
Image CRADIO v4-H BF16 (kept full-precision in every non-GGUF variant; GGUF uses mmproj-F16 split file)
Audio Parakeet-TDT-0.6B-v2 BF16 (same rationale)
Video Parakeet-TDT-0.6B-v2 + frame sampler BF16 (≤ 2 min, 256 frames @ 2 FPS)

NVIDIA's official FP8 / NVFP4 recipe keeps both encoders + the cross-modal MLP projectors in BF16 to preserve multimodal accuracy. We follow that convention in every quantized variant we ship.

Runtime quirks

MLX-LM (text-only)

This variant covers the LLM backbone only. Vision + audio encoders are NOT included — MLX-VLM Nemotron-Omni model class is pending upstream support (no PR observed as of 2026-05-04).

Use the mlx_lm.generate API; enable_thinking is a runtime flag (see below).

Reasoning mode

enable_thinking defaults to True. To disable extended reasoning (e.g., for latency-sensitive cases), pass enable_thinking=False to the chat template / generate call. No separate "no-think" variant card exists — this is a runtime flag, not a model variant.