NVIDIA

Nemotron Mini 4B

Name: Nemotron Mini 4B
Rating: 50 (193 reviews)
Author: NVIDIA

当前

HuggingFace

376.4K下载量182点赞Aug 2024发布日期4K tokens上下文NVIDIA Open Model许可证6 入门质量

Nemotron Mini 4B (4B parameters) requires approximately 6.2 GB of VRAM with Q4_K_M quantization. For the best balance of quality and speed, we recommend hardware with at least 8 GB of VRAM.

快速开始

— 复制粘贴即可本地运行

Copy-paste commands to run Nemotron Mini 4B on your machine.

Run

docker run --rm -it ghcr.io/ggerganov/llama.cpp:full \
  --hf-repo "nvidia/Nemotron-Mini-4B-Instruct" \
  --hf-file "Nemotron-Mini-4B-Instruct-Q4_K_M.gguf" \
  -c 4096 -ngl 99

Quick specs

Parameters4B

Architecturedense

Context4K tokens

Modalitytext

Min RAM1.6 GB

Rec. RAM2.4 GB (Q4_K_M)

LicenseNVIDIA Open Model

FamilyNemotron

✓ Chat

About this model

Nemotron-Mini-4B-Instruct is a model for generating responses for roleplaying, retrieval augmented generation, and function calling. It is a small language model (SLM) optimized through distillation, pruning and quantization for speed and on-device deployment. It is a fine-tuned version of nvidia/Minitron-4B-Base, which was pruned and distilled from Nemotron-4 15B using our LLM compression technique. This instruct model is optimized for roleplay, RAG QA, and function calling in English. It supports a context length of 4,096 tokens. This model is ready for commercial use.

•Garak, is an automated LLM vulnerability scanner that probes for common weaknesses, including prompt injection and data leakage
•AEGIS, is a content safety evaluation dataset and LLM based content safety classifier model, that adheres to a broad taxonomy of 13 categories of...
•Human Content Red Teaming leveraging human interaction and evaluation of the models' responses

Nemotron Mini 4B 的最佳选择

运行此模型

Nemotron Mini 4B on RTX 5060 8GB Nemotron Mini 4B on RTX 5060 Ti 8GB Nemotron Mini 4B on RTX 5050 8GB

量化选项

各量化级别的 VRAM 估算

No hardware detected — fit column shows raw VRAM estimates

Quant	Bits	VRAM	Quality	Fit
Q2_K	2	1.6 GB	Low	—
Q3_K_S	3	2.0 GB	Low	—
NVFP4	4	2.2 GB	Medium	—
Q4_K_M	4	2.4 GB	Medium	—
Q5_K_M	5	2.9 GB	High	—
Q6_K	6	3.3 GB	High	—
Q8_0	8	4.3 GB	Very High	—
F16	16	8.2 GB	Maximum	—

Quality benchmarks

Nemotron Mini 4B benchmark scores

Benchmark verified

Coding

SWE-bench Verified—

HumanEval+23.3%

Aider Polyglot—

LiveCodeBench—

Reasoning

MMLU-Pro18.1%

GPQA Diamond4.0%

MATH-5002.6%

ARC Challenge50.9%

General

Chatbot Arena—

IFEval66.7%

Source: official · 2024-07-17

硬件兼容性

全部硬件的适配估算

打开计算器

Computing compatibility...

内存详细分析

Reference: RTX 2060 6GB

Weights2.4 GB

KV Cache2.0 GB

Runtime1.2 GB

Headroom0.6 GB

常见问题

FAQ — Nemotron Mini 4B

另请参阅

量化指南评分方法打开计算器