Nemotron-H-47B-Base-8K

Home/Models/nvidia/Nemotron-H-47B-Base-8K

nvidia/Nemotron-H-47B-Base-8K

kind: original
quant: bf16 (16-bit)
family: nemotron
modality: text-generation
root model: nvidia/Nemotron-H-56B-Base-8K
provider: nvidia
license: other
language: en, de, es, fr, it, ko, pt, ru, jp, zh
context: 8K
library: transformers
vRAM (GB): —
size on disk: —
downloads: 271
likes: 22
created: —
last modified: 2025-08-21T15:45:32+00:00 (2025)
huggingface: https://huggingface.co/nvidia/Nemotron-H-47B-Base-8K
tags: transformerssafetensorsnvidiapytorchnemotron-htext-generationendeesfritkoptrujpzharxiv:2504.03624base_model:nvidia/Nemotron-H-56B-Base-8Kbase_model:finetune:nvidia/Nemotron-H-56B-Base-8Klicense:otherendpoints_compatibleregion:us

benchmark	category	score	context	conf	evidence
GSM8K	math	93.3	GSM8K 8-shot CoT \| MATH 4-shot CoT \| MATH-Lvl 5 4-shot CoT \| MATH-500 4-shot CoT	low	`\| 93.3 \| 57.4 \| 34.2 \| 57.9 \|`
CommonsenseQA	knowledge	87.3		low	`\| 94.6 \| 87.9 \| 83.9 \| 87.3 \|`
WinoGrande	knowledge	83.9		low	`\| 94.6 \| 87.9 \| 83.9 \| 87.3 \|`
MMLU	knowledge	83.6	MMLU-Pro 5-shot-cot	low	`\| 61.8 \| 83.6 \|`
MBPP	coding	75.9		low	`\| 75.9 \| 65.6 \| 61.0 \| 56.1 \|`
MMLU-Pro	knowledge	61.8	MMLU-Pro 5-shot-cot	low	`\| 61.8 \| 83.6 \|`
HumanEval	coding	61.0		low	`\| 75.9 \| 65.6 \| 61.0 \| 56.1 \|`
MATH-500	math	57.9	GSM8K 8-shot CoT \| MATH 4-shot CoT \| MATH-Lvl 5 4-shot CoT \| MATH-500 4-shot CoT	low	`\| 93.3 \| 57.4 \| 34.2 \| 57.9 \|`
MATH	math	57.4	GSM8K 8-shot CoT \| MATH 4-shot CoT \| MATH-Lvl 5 4-shot CoT \| MATH-500 4-shot CoT	low	`\| 93.3 \| 57.4 \| 34.2 \| 57.9 \|`
ARC-Challenge	knowledge	25.0		medium	`ARC Challenge 25`
HellaSwag	knowledge	10.0		medium	`Hellaswag 10`

none scraped.