Nemotron-H-4B-Base-8K

Home/Models/nvidia/Nemotron-H-4B-Base-8K

nvidia/Nemotron-H-4B-Base-8K

kind: original
quant: bf16 (16-bit)
family: nemotron
modality: text-generation
root model: nvidia/Nemotron-H-4B-Base-8K
provider: nvidia
license: other
language: en
context: 8K
library: transformers
vRAM (GB): —
size on disk: —
downloads: 3,762
likes: 7
created: —
last modified: 2025-10-24T00:42:23+00:00 (2025)
huggingface: https://huggingface.co/nvidia/Nemotron-H-4B-Base-8K
tags: transformerssafetensorsnvidiapytorchtext-generationenarxiv:2504.11409license:otherendpoints_compatibleregion:us

benchmark	category	score	conf	evidence
HellaSwag	knowledge	77.0	low	`\| 54.4 \| 77.0 \| 71.3 \| 70.2 \|`
WinoGrande	knowledge	71.3	low	`\| 54.4 \| 77.0 \| 71.3 \| 70.2 \|`
CommonsenseQA	knowledge	70.2	low	`\| 54.4 \| 77.0 \| 71.3 \| 70.2 \|`
MBPP	coding	65.0	low	`\| 65.0 \| 61.1 \| 59.8 \| 55.5 \|`
HumanEval	coding	59.8	low	`\| 65.0 \| 61.1 \| 59.8 \| 55.5 \|`
ARC-Challenge	knowledge	54.4	low	`\| 54.4 \| 77.0 \| 71.3 \| 70.2 \|`

none scraped.