jkim96

Nemotron-3-Ultra-550B-A55B-DASHQ-INT2-g32

Dedicated Endpoints

Run this model inference on single tenant GPU with unmatched speed and reliability at scale.

Get help setting up a custom Dedicated Endpoints.

Talk with our engineer to get a quote for reserved GPU instances with discounts.

License: other

bash
pip install git+https://github.com/JaeminK/dashq.git

python
from dashq import load_quantized

model, tokenizer = load_quantized("jkim96/Nemotron-3-Ultra-550B-A55B-DASHQ-INT2-g32", device_map="auto")

Table
Field	Value
Base model	`nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16`
Precision	INT2, group size 32
Scale / zero dtype	float16
Calibration	wikitext2, 128 samples x 2048
Size	209.8371 GB · original 1121.0559 GB · 5.3x compression

Full zero-shot / few-shot results for every DASH-Q checkpoint: github.com/JaeminK/dashq#benchmarks

Model provider

jkim96

Model tree

Base

nvidia/NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16

Quantized

this model

Modalities

Input

Text

Output

Text

Pricing

Dedicated Endpoints

Supported Functionality

Model APIs

Dedicated Endpoints

Container

More information