142 lines
6.4 KiB
Markdown
142 lines
6.4 KiB
Markdown
---
|
||
license: other
|
||
license_name: nvidia-open-model-license
|
||
license_link: >-
|
||
https://www.nvidia.com/en-us/agreements/enterprise-software/nvidia-open-model-license/
|
||
base_model: nvidia/NVIDIA-Nemotron-Nano-9B-v2
|
||
model_creator: nvidia
|
||
model_name: NVIDIA-Nemotron-Nano-9B-v2
|
||
quantized_by: Second State Inc.
|
||
pipeline_tag: text-generation
|
||
language:
|
||
- en
|
||
- es
|
||
- fr
|
||
- de
|
||
- it
|
||
- ja
|
||
library_name: transformers
|
||
---
|
||
|
||
<!-- header start -->
|
||
<!-- 200823 -->
|
||
<div style="width: auto; margin-left: auto; margin-right: auto">
|
||
<img src="https://github.com/LlamaEdge/LlamaEdge/raw/dev/assets/logo.svg" style="width: 100%; min-width: 400px; display: block; margin: auto;">
|
||
</div>
|
||
<hr style="margin-top: 1.0em; margin-bottom: 1.0em;">
|
||
<!-- header end -->
|
||
|
||
# NVIDIA-Nemotron-Nano-9B-v2-GGUF
|
||
|
||
## Original Model
|
||
|
||
[nvidia/NVIDIA-Nemotron-Nano-9B-v2](https://huggingface.co/nvidia/NVIDIA-Nemotron-Nano-9B-v2)
|
||
|
||
## Run with LlamaEdge
|
||
|
||
- LlamaEdge version: coming soon
|
||
|
||
<!-- - LlamaEdge version: [v0.25.1](https://github.com/LlamaEdge/LlamaEdge/releases/tag/0.25.1) and above -->
|
||
|
||
- Prompt template
|
||
|
||
- Prompt type: `nemotron-2-chat`
|
||
|
||
- Prompt string
|
||
- `Thinking` mode
|
||
|
||
```text
|
||
<SPECIAL_10>System
|
||
You are a helpful assistant.
|
||
|
||
<SPECIAL_11>User
|
||
Hi, can you explain what quantum computing is? /think
|
||
|
||
<SPECIAL_11>Assistant
|
||
<think>
|
||
The user is asking about quantum computing, I should give a simple explanation.
|
||
</think>
|
||
Quantum computing uses principles of quantum mechanics, such as superposition and entanglement, to perform computations that would be hard for classical computers.
|
||
<SPECIAL_12>
|
||
|
||
<SPECIAL_11>User
|
||
That sounds complicated. Can you give me an example? /think
|
||
|
||
<SPECIAL_11>Assistant
|
||
<think>
|
||
I should give a concrete example, like Shor’s algorithm or Grover’s algorithm, but in simple terms.
|
||
</think>
|
||
For example, a quantum computer can factor very large numbers much faster than classical computers, which could break some encryption methods.
|
||
<SPECIAL_12>
|
||
|
||
<SPECIAL_11>User
|
||
Okay, explain it in a simpler way for a 10-year-old. /think
|
||
|
||
<SPECIAL_11>Assistant
|
||
<think>
|
||
```
|
||
|
||
- `No-thinking` mode
|
||
|
||
```text
|
||
<SPECIAL_10>System
|
||
You are a helpful assistant.
|
||
|
||
/no_think
|
||
|
||
<SPECIAL_11>User
|
||
Hi, can you explain what quantum computing is? /no_think
|
||
|
||
<SPECIAL_11>Assistant
|
||
<think></think>
|
||
Quantum computing is a type of computing that uses the rules of quantum mechanics to solve certain problems faster than regular computers.
|
||
<SPECIAL_12>
|
||
|
||
<SPECIAL_11>User
|
||
That sounds complicated. Can you give me an example? /no_think
|
||
|
||
<SPECIAL_11>Assistant
|
||
<think></think>
|
||
For example, quantum computers could quickly factor very large numbers, which is important for cryptography.
|
||
<SPECIAL_12>
|
||
|
||
<SPECIAL_11>User
|
||
Okay, explain it even more simply. /no_think
|
||
|
||
<SPECIAL_11>Assistant
|
||
<think></think>
|
||
```
|
||
|
||
- Context size: `128000`
|
||
|
||
- Run as LlamaEdge service
|
||
|
||
```bash
|
||
wasmedge --dir .:. \
|
||
--nn-preload default:GGML:AUTO:NVIDIA-Nemotron-Nano-9B-v2-Q5_K_M.gguf \
|
||
llama-api-server.wasm \
|
||
--prompt-template nemotron-2-chat \
|
||
--ctx-size 128000 \
|
||
--model-name nemotron-nano-v2
|
||
```
|
||
|
||
## Quantized GGUF Models
|
||
|
||
| Name | Quant method | Bits | Size | Use case |
|
||
| ---- | ---- | ---- | ---- | ----- |
|
||
| [NVIDIA-Nemotron-Nano-9B-v2-Q2_K.gguf](https://huggingface.co/second-state/NVIDIA-Nemotron-Nano-9B-v2-GGUF/blob/main/NVIDIA-Nemotron-Nano-9B-v2-Q2_K.gguf) | Q2_K | 2 | 5.01 GB| smallest, significant quality loss - not recommended for most purposes |
|
||
| [NVIDIA-Nemotron-Nano-9B-v2-Q3_K_L.gguf](https://huggingface.co/second-state/NVIDIA-Nemotron-Nano-9B-v2-GGUF/blob/main/NVIDIA-Nemotron-Nano-9B-v2-Q3_K_L.gguf) | Q3_K_L | 3 | 5.49 GB| small, substantial quality loss |
|
||
| [NVIDIA-Nemotron-Nano-9B-v2-Q3_K_M.gguf](https://huggingface.co/second-state/NVIDIA-Nemotron-Nano-9B-v2-GGUF/blob/main/NVIDIA-Nemotron-Nano-9B-v2-Q3_K_M.gguf) | Q3_K_M | 3 | 5.38 GB| very small, high quality loss |
|
||
| [NVIDIA-Nemotron-Nano-9B-v2-Q3_K_S.gguf](https://huggingface.co/second-state/NVIDIA-Nemotron-Nano-9B-v2-GGUF/blob/main/NVIDIA-Nemotron-Nano-9B-v2-Q3_K_S.gguf) | Q3_K_S | 3 | 5.13 GB| very small, high quality loss |
|
||
| [NVIDIA-Nemotron-Nano-9B-v2-Q4_0.gguf](https://huggingface.co/second-state/NVIDIA-Nemotron-Nano-9B-v2-GGUF/blob/main/NVIDIA-Nemotron-Nano-9B-v2-Q4_0.gguf) | Q4_0 | 4 | 5.31 GB| legacy; small, very high quality loss - prefer using Q3_K_M |
|
||
| [NVIDIA-Nemotron-Nano-9B-v2-Q4_K_M.gguf](https://huggingface.co/second-state/NVIDIA-Nemotron-Nano-9B-v2-GGUF/blob/main/NVIDIA-Nemotron-Nano-9B-v2-Q4_K_M.gguf) | Q4_K_M | 4 | 6.53 GB| medium, balanced quality - recommended |
|
||
| [NVIDIA-Nemotron-Nano-9B-v2-Q4_K_S.gguf](https://huggingface.co/second-state/NVIDIA-Nemotron-Nano-9B-v2-GGUF/blob/main/NVIDIA-Nemotron-Nano-9B-v2-Q4_K_S.gguf) | Q4_K_S | 4 | 6.21 GB| small, greater quality loss |
|
||
| [NVIDIA-Nemotron-Nano-9B-v2-Q5_0.gguf](https://huggingface.co/second-state/NVIDIA-Nemotron-Nano-9B-v2-GGUF/blob/main/NVIDIA-Nemotron-Nano-9B-v2-Q5_0.gguf) | Q5_0 | 5 | 6.35 GB| legacy; medium, balanced quality - prefer using Q4_K_M |
|
||
| [NVIDIA-Nemotron-Nano-9B-v2-Q5_K_M.gguf](https://huggingface.co/second-state/NVIDIA-Nemotron-Nano-9B-v2-GGUF/blob/main/NVIDIA-Nemotron-Nano-9B-v2-Q5_K_M.gguf) | Q5_K_M | 5 | 7.07 GB| large, very low quality loss - recommended |
|
||
| [NVIDIA-Nemotron-Nano-9B-v2-Q5_K_S.gguf](https://huggingface.co/second-state/NVIDIA-Nemotron-Nano-9B-v2-GGUF/blob/main/NVIDIA-Nemotron-Nano-9B-v2-Q5_K_S.gguf) | Q5_K_S | 5 | 6.78 GB| large, low quality loss - recommended |
|
||
| [NVIDIA-Nemotron-Nano-9B-v2-Q6_K.gguf](https://huggingface.co/second-state/NVIDIA-Nemotron-Nano-9B-v2-GGUF/blob/main/NVIDIA-Nemotron-Nano-9B-v2-Q6_K.gguf) | Q6_K | 6 | 9.14 GB| very large, extremely low quality loss |
|
||
| [NVIDIA-Nemotron-Nano-9B-v2-Q8_0.gguf](https://huggingface.co/second-state/Nemotron-Mini-4B-Instruct-GGUF/blob/main/NVIDIA-Nemotron-Nano-9B-v2-f16.gguf) | Q8_0 | 8 | 17.8 GB| very large, extremely low quality loss - not recommended |
|
||
| [NVIDIA-Nemotron-Nano-9B-v2-f16.gguf](https://huggingface.co/second-state/NVIDIA-Nemotron-Nano-9B-v2-GGUF/blob/main/NVIDIA-Nemotron-Nano-9B-v2-f16-00001-of-00003.gguf) | f16 | 16 | 30.0 GB| |
|
||
|
||
*Quantized with llama.cpp b6315.*
|