---
base_model: Nanbeige/Nanbeige-16B-Base-32K
inference: false
language:
- en
- zh
library_name: transformers
license: apache-2.0
model_creator: Nanbeige LLM Lab
model_name: Nanbeige 16B Base 32K
model_type: nanbeige
pipeline_tag: text-generation
prompt_template: '{prompt}
'
quantized_by: TheBloke
tags:
- llm
- nanbeige
- custom_code
---
# Nanbeige 16B Base 32K - GGUF
- Model creator: [Nanbeige LLM Lab](https://huggingface.co/Nanbeige)
- Original model: [Nanbeige 16B Base 32K](https://huggingface.co/Nanbeige/Nanbeige-16B-Base-32K)
## Description
This repo contains GGUF format model files for [Nanbeige LLM Lab's Nanbeige 16B Base 32K](https://huggingface.co/Nanbeige/Nanbeige-16B-Base-32K).
These files were quantised using hardware kindly provided by [Massed Compute](https://massedcompute.com/).
### About GGUF
GGUF is a new format introduced by the llama.cpp team on August 21st 2023. It is a replacement for GGML, which is no longer supported by llama.cpp.
Here is an incomplete list of clients and libraries that are known to support GGUF:
* [llama.cpp](https://github.com/ggerganov/llama.cpp). The source project for GGUF. Offers a CLI and a server option.
* [text-generation-webui](https://github.com/oobabooga/text-generation-webui), the most widely used web UI, with many features and powerful extensions. Supports GPU acceleration.
* [KoboldCpp](https://github.com/LostRuins/koboldcpp), a fully featured web UI, with GPU accel across all platforms and GPU architectures. Especially good for story telling.
* [LM Studio](https://lmstudio.ai/), an easy-to-use and powerful local GUI for Windows and macOS (Silicon), with GPU acceleration.
* [LoLLMS Web UI](https://github.com/ParisNeo/lollms-webui), a great web UI with many interesting and unique features, including a full model library for easy model selection.
* [Faraday.dev](https://faraday.dev/), an attractive and easy to use character-based chat GUI for Windows and macOS (both Silicon and Intel), with GPU acceleration.
* [ctransformers](https://github.com/marella/ctransformers), a Python library with GPU accel, LangChain support, and OpenAI-compatible AI server.
* [llama-cpp-python](https://github.com/abetlen/llama-cpp-python), a Python library with GPU accel, LangChain support, and OpenAI-compatible API server.
* [candle](https://github.com/huggingface/candle), a Rust ML framework with a focus on performance, including GPU support, and ease of use.
## Repositories available
* [GPTQ models for GPU inference, with multiple quantisation parameter options.](https://huggingface.co/TheBloke/Nanbeige-16B-Base-32K-GPTQ)
* [2, 3, 4, 5, 6 and 8-bit GGUF models for CPU+GPU inference](https://huggingface.co/TheBloke/Nanbeige-16B-Base-32K-GGUF)
* [Nanbeige LLM Lab's original unquantised fp16 model in pytorch format, for GPU inference and for further conversions](https://huggingface.co/Nanbeige/Nanbeige-16B-Base-32K)
## Prompt template: None
```
{prompt}
```
## Compatibility
These quantised GGUFv2 files are compatible with llama.cpp from August 27th onwards, as of commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221)
They are also compatible with many third party UIs and libraries - please see the list at the top of this README.
## Explanation of quantisation methods
Click to see details
The new methods available are:
* GGML_TYPE_Q2_K - "type-1" 2-bit quantization in super-blocks containing 16 blocks, each block having 16 weight. Block scales and mins are quantized with 4 bits. This ends up effectively using 2.5625 bits per weight (bpw)
* GGML_TYPE_Q3_K - "type-0" 3-bit quantization in super-blocks containing 16 blocks, each block having 16 weights. Scales are quantized with 6 bits. This end up using 3.4375 bpw.
* GGML_TYPE_Q4_K - "type-1" 4-bit quantization in super-blocks containing 8 blocks, each block having 32 weights. Scales and mins are quantized with 6 bits. This ends up using 4.5 bpw.
* GGML_TYPE_Q5_K - "type-1" 5-bit quantization. Same super-block structure as GGML_TYPE_Q4_K resulting in 5.5 bpw
* GGML_TYPE_Q6_K - "type-0" 6-bit quantization. Super-blocks with 16 blocks, each block having 16 weights. Scales are quantized with 8 bits. This ends up using 6.5625 bpw
Refer to the Provided Files table below to see what files use which methods, and how.
## Provided files
| Name | Quant method | Bits | Size | Max RAM required | Use case |
| ---- | ---- | ---- | ---- | ---- | ----- |
| [nanbeige-16b-base-32k.Q2_K.gguf](https://huggingface.co/TheBloke/Nanbeige-16B-Base-32K-GGUF/blob/main/nanbeige-16b-base-32k.Q2_K.gguf) | Q2_K | 2 | 6.64 GB| 9.14 GB | smallest, significant quality loss - not recommended for most purposes |
| [nanbeige-16b-base-32k.Q3_K_S.gguf](https://huggingface.co/TheBloke/Nanbeige-16B-Base-32K-GGUF/blob/main/nanbeige-16b-base-32k.Q3_K_S.gguf) | Q3_K_S | 3 | 6.93 GB| 9.43 GB | very small, high quality loss |
| [nanbeige-16b-base-32k.Q3_K_M.gguf](https://huggingface.co/TheBloke/Nanbeige-16B-Base-32K-GGUF/blob/main/nanbeige-16b-base-32k.Q3_K_M.gguf) | Q3_K_M | 3 | 7.74 GB| 10.24 GB | very small, high quality loss |
| [nanbeige-16b-base-32k.Q3_K_L.gguf](https://huggingface.co/TheBloke/Nanbeige-16B-Base-32K-GGUF/blob/main/nanbeige-16b-base-32k.Q3_K_L.gguf) | Q3_K_L | 3 | 8.45 GB| 10.95 GB | small, substantial quality loss |
| [nanbeige-16b-base-32k.Q4_0.gguf](https://huggingface.co/TheBloke/Nanbeige-16B-Base-32K-GGUF/blob/main/nanbeige-16b-base-32k.Q4_0.gguf) | Q4_0 | 4 | 8.99 GB| 11.49 GB | legacy; small, very high quality loss - prefer using Q3_K_M |
| [nanbeige-16b-base-32k.Q4_K_S.gguf](https://huggingface.co/TheBloke/Nanbeige-16B-Base-32K-GGUF/blob/main/nanbeige-16b-base-32k.Q4_K_S.gguf) | Q4_K_S | 4 | 9.04 GB| 11.54 GB | small, greater quality loss |
| [nanbeige-16b-base-32k.Q4_K_M.gguf](https://huggingface.co/TheBloke/Nanbeige-16B-Base-32K-GGUF/blob/main/nanbeige-16b-base-32k.Q4_K_M.gguf) | Q4_K_M | 4 | 9.59 GB| 12.09 GB | medium, balanced quality - recommended |
| [nanbeige-16b-base-32k.Q5_0.gguf](https://huggingface.co/TheBloke/Nanbeige-16B-Base-32K-GGUF/blob/main/nanbeige-16b-base-32k.Q5_0.gguf) | Q5_0 | 5 | 10.93 GB| 13.43 GB | legacy; medium, balanced quality - prefer using Q4_K_M |
| [nanbeige-16b-base-32k.Q5_K_S.gguf](https://huggingface.co/TheBloke/Nanbeige-16B-Base-32K-GGUF/blob/main/nanbeige-16b-base-32k.Q5_K_S.gguf) | Q5_K_S | 5 | 10.93 GB| 13.43 GB | large, low quality loss - recommended |
| [nanbeige-16b-base-32k.Q5_K_M.gguf](https://huggingface.co/TheBloke/Nanbeige-16B-Base-32K-GGUF/blob/main/nanbeige-16b-base-32k.Q5_K_M.gguf) | Q5_K_M | 5 | 11.24 GB| 13.74 GB | large, very low quality loss - recommended |
| [nanbeige-16b-base-32k.Q6_K.gguf](https://huggingface.co/TheBloke/Nanbeige-16B-Base-32K-GGUF/blob/main/nanbeige-16b-base-32k.Q6_K.gguf) | Q6_K | 6 | 12.99 GB| 15.49 GB | very large, extremely low quality loss |
| [nanbeige-16b-base-32k.Q8_0.gguf](https://huggingface.co/TheBloke/Nanbeige-16B-Base-32K-GGUF/blob/main/nanbeige-16b-base-32k.Q8_0.gguf) | Q8_0 | 8 | 16.83 GB| 19.33 GB | very large, extremely low quality loss - not recommended |
**Note**: the above RAM figures assume no GPU offloading. If layers are offloaded to the GPU, this will reduce RAM usage and use VRAM instead.
## How to download GGUF files
**Note for manual downloaders:** You almost never want to clone the entire repo! Multiple different quantisation formats are provided, and most users only want to pick and download a single file.
The following clients/libraries will automatically download models for you, providing a list of available models to choose from:
* LM Studio
* LoLLMS Web UI
* Faraday.dev
### In `text-generation-webui`
Under Download Model, you can enter the model repo: TheBloke/Nanbeige-16B-Base-32K-GGUF and below it, a specific filename to download, such as: nanbeige-16b-base-32k.Q4_K_M.gguf.
Then click Download.
### On the command line, including multiple files at once
I recommend using the `huggingface-hub` Python library:
```shell
pip3 install huggingface-hub
```
Then you can download any individual model file to the current directory, at high speed, with a command like this:
```shell
huggingface-cli download TheBloke/Nanbeige-16B-Base-32K-GGUF nanbeige-16b-base-32k.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False
```
More advanced huggingface-cli download usage
You can also download multiple files at once with a pattern:
```shell
huggingface-cli download TheBloke/Nanbeige-16B-Base-32K-GGUF --local-dir . --local-dir-use-symlinks False --include='*Q4_K*gguf'
```
For more documentation on downloading with `huggingface-cli`, please see: [HF -> Hub Python Library -> Download files -> Download from the CLI](https://huggingface.co/docs/huggingface_hub/guides/download#download-from-the-cli).
To accelerate downloads on fast connections (1Gbit/s or higher), install `hf_transfer`:
```shell
pip3 install hf_transfer
```
And set environment variable `HF_HUB_ENABLE_HF_TRANSFER` to `1`:
```shell
HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli download TheBloke/Nanbeige-16B-Base-32K-GGUF nanbeige-16b-base-32k.Q4_K_M.gguf --local-dir . --local-dir-use-symlinks False
```
Windows Command Line users: You can set the environment variable by running `set HF_HUB_ENABLE_HF_TRANSFER=1` before the download command.
## Example `llama.cpp` command
Make sure you are using `llama.cpp` from commit [d0cee0d](https://github.com/ggerganov/llama.cpp/commit/d0cee0d36d5be95a0d9088b674dbb27354107221) or later.
```shell
./main -ngl 32 -m nanbeige-16b-base-32k.Q4_K_M.gguf --color -c 2048 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "{prompt}"
```
Change `-ngl 32` to the number of layers to offload to GPU. Remove it if you don't have GPU acceleration.
Change `-c 2048` to the desired sequence length. For extended sequence models - eg 8K, 16K, 32K - the necessary RoPE scaling parameters are read from the GGUF file and set by llama.cpp automatically.
If you want to have a chat-style conversation, replace the `-p ` argument with `-i -ins`
For other parameters and how to use them, please refer to [the llama.cpp documentation](https://github.com/ggerganov/llama.cpp/blob/master/examples/main/README.md)
## How to run in `text-generation-webui`
Further instructions can be found in the text-generation-webui documentation, here: [text-generation-webui/docs/04 ‐ Model Tab.md](https://github.com/oobabooga/text-generation-webui/blob/main/docs/04%20%E2%80%90%20Model%20Tab.md#llamacpp).
## How to run from Python code
You can use GGUF models from Python using the [llama-cpp-python](https://github.com/abetlen/llama-cpp-python) or [ctransformers](https://github.com/marella/ctransformers) libraries.
### How to load this model in Python code, using ctransformers
#### First install the package
Run one of the following commands, according to your system:
```shell
# Base ctransformers with no GPU acceleration
pip install ctransformers
# Or with CUDA GPU acceleration
pip install ctransformers[cuda]
# Or with AMD ROCm GPU acceleration (Linux only)
CT_HIPBLAS=1 pip install ctransformers --no-binary ctransformers
# Or with Metal GPU acceleration for macOS systems only
CT_METAL=1 pip install ctransformers --no-binary ctransformers
```
#### Simple ctransformers example code
```python
from ctransformers import AutoModelForCausalLM
# Set gpu_layers to the number of layers to offload to GPU. Set to 0 if no GPU acceleration is available on your system.
llm = AutoModelForCausalLM.from_pretrained("TheBloke/Nanbeige-16B-Base-32K-GGUF", model_file="nanbeige-16b-base-32k.Q4_K_M.gguf", model_type="nanbeige", gpu_layers=50)
print(llm("AI is going to"))
```
## How to use with LangChain
Here are guides on using llama-cpp-python and ctransformers with LangChain:
* [LangChain + llama-cpp-python](https://python.langchain.com/docs/integrations/llms/llamacpp)
* [LangChain + ctransformers](https://python.langchain.com/docs/integrations/providers/ctransformers)
## Discord
For further support, and discussions on these models and AI in general, join us at:
[TheBloke AI's Discord server](https://discord.gg/theblokeai)
## Thanks, and how to contribute
Thanks to the [chirper.ai](https://chirper.ai) team!
Thanks to Clay from [gpus.llm-utils.org](llm-utils)!
I've had a lot of people ask if they can contribute. I enjoy providing models and helping people, and would love to be able to spend even more time doing it, as well as expanding into new projects like fine tuning/training.
If you're able and willing to contribute it will be most gratefully received and will help me to keep providing more models, and to start work on new AI projects.
Donaters will get priority support on any and all AI/LLM/model questions and requests, access to a private Discord room, plus other benefits.
* Patreon: https://patreon.com/TheBlokeAI
* Ko-Fi: https://ko-fi.com/TheBlokeAI
**Special thanks to**: Aemon Algiz.
**Patreon special mentions**: Brandon Frisco, LangChain4j, Spiking Neurons AB, transmissions 11, Joseph William Delisle, Nitin Borwankar, Willem Michiel, Michael Dempsey, vamX, Jeffrey Morgan, zynix, jjj, Omer Bin Jawed, Sean Connelly, jinyuan sun, Jeromy Smith, Shadi, Pawan Osman, Chadd, Elijah Stavena, Illia Dulskyi, Sebastain Graf, Stephen Murray, terasurfer, Edmond Seymore, Celu Ramasamy, Mandus, Alex, biorpg, Ajan Kanaga, Clay Pascal, Raven Klaugh, 阿明, K, ya boyyy, usrbinkat, Alicia Loh, John Villwock, ReadyPlayerEmma, Chris Smitley, Cap'n Zoog, fincy, GodLy, S_X, sidney chen, Cory Kujawski, OG, Mano Prime, AzureBlack, Pieter, Kalila, Spencer Kim, Tom X Nguyen, Stanislav Ovsiannikov, Michael Levine, Andrey, Trailburnt, Vadim, Enrico Ros, Talal Aujan, Brandon Phillips, Jack West, Eugene Pentland, Michael Davis, Will Dee, webtim, Jonathan Leane, Alps Aficionado, Rooh Singh, Tiffany J. Kim, theTransient, Luke @flexchar, Elle, Caitlyn Gatomon, Ari Malik, subjectnull, Johann-Peter Hartmann, Trenton Dambrowitz, Imad Khwaja, Asp the Wyvern, Emad Mostaque, Rainer Wilmers, Alexandros Triantafyllidis, Nicholas, Pedro Madruga, SuperWojo, Harry Royden McLaughlin, James Bentley, Olakabola, David Ziegler, Ai Maven, Jeff Scroggin, Nikolai Manek, Deo Leter, Matthew Berman, Fen Risland, Ken Nordquist, Manuel Alberto Morcote, Luke Pendergrass, TL, Fred von Graf, Randy H, Dan Guido, NimbleBox.ai, Vitor Caleffi, Gabriel Tamborski, knownsqashed, Lone Striker, Erik Bjäreholt, John Detwiler, Leonard Tan, Iucharbius
Thank you to all my generous patrons and donaters!
And thank you again to a16z for their generous grant.
# Original model card: Nanbeige LLM Lab's Nanbeige 16B Base 32K
Nanbeige-16B-Base-32k
💻Github
# 模型介绍（Introduction）
Nanbeige-16B（南北阁-16B）是南北阁大模型实验室研发的160亿参数规模的大语言模型，采用了2.5T Tokens进行预训练，数据包含大量互联网高质量语料、各类书籍、代码等领域脱敏文本，在各个权威测评数据集上都取得了不错的效果。本次发布包含有 Base、Chat 以及扩展上下文长度的 Base-32k、Chat-32k 版本。
Base-32k 版本基于Nanbeige-16B-Base模型，采用YaRN插值方法对位置编码进行修改，并以32k为最大长度进行了20B Tokens的中文、英文、代码语料的全参数增量预训练。
Chat 版本和 Chat-32k 版本分别基于Nanbeige-16B-Base模型和Nanbeige-16B-Base-32k模型，经过了大量人类对齐训练，能够更好、更安全地回复用户的问题。
如果您需要处理更长的上下文，我们推荐您使用Nanbeige-16B-Base-32k和Nanbeige-16B-Chat-32k。
本仓库为 Nanbeige-16B-Base-32k 的模型仓库。
Nanbeige-16B is a 16 billion parameter language model developed by Nanbeige LLM Lab. It uses 2.5T Tokens for pre-training. The training data includes a large amount of high-quality internet corpus, various books, code, etc. It has achieved good results on various authoritative evaluation data sets. This release includes the Base, Chat, Base-32k and Chat-32k.
The Base-32k version is based on the Nanbeige-16B-Base model, which uses YaRN interpolation method to modify the position encoding, and performs full parameter incremental pre-training with 20 billion tokens of Chinese, English, and code corpora, under 32k max length.
The Chat version and Chat-32k version are based on the Nanbeige-16B-Base model and Nanbeige-16B-Base-32k model respectively. They have undergone extensive human-aligned training.
If you need to deal with longer contexts, we recommend using Nanbeige-16B-Base-32k and Nanbeige-16B-Chat-32k.
This repository is the one for Nanbeige-16B-Base-32k model.
##
| | Base Model | Base-32k Model | Chat Model | Chat-32k Model |
|:-------:|:-------:|:-------:|:---------:|:--------:|
| 16B | 🤗 [Nanbeige-16B-Base](https://huggingface.co/Nanbeige/Nanbeige-16B-Base) | 🤗 [Nanbeige-16B-Base-32k](https://huggingface.co/Nanbeige/Nanbeige-16B-Base-32k) | 🤗 [Nanbeige-16B-Chat](https://huggingface.co/Nanbeige/Nanbeige-16B-Chat) |🤗 [Nanbeige-16B-Chat-32k](https://huggingface.co/Nanbeige/Nanbeige-16B-Chat-32k) |
##
# 模型推理 (Inference)
## 相关依赖
- python 3.8及以上版本
- transformers 4.33.3
- pytorch 2.0及以上版本
- python 3.8 and above
- transformers 4.33.3
- pytorch 2.0及以上版本
- deepspeed 0.11.1及以上版本
可以通过以下pip命令安装相关依赖库
You can install the dependent libraries with the following pip command
```
pip install transformers==4.33.3 transformers_stream_generator deepspeed einops==0.3.2 datasets==2.10.0 deepspeed==0.11.1
```
## 推理代码
通过以下代码可以调用模型进行续写生成：
The following code can be used to invoke the model for text generation:
```python
import torch
from transformers import AutoModelForCausalLM, AutoTokenizer
from transformers.generation.utils import GenerationConfig
import deepspeed
import os
tokenizer = AutoTokenizer.from_pretrained("Nanbeige/Nanbeige-16B-Base-32k", use_fast=False, trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("Nanbeige/Nanbeige-16B-Base-32k", device_map="auto", torch_dtype=torch.bfloat16, trust_remote_code=True)
world_size = int(os.getenv('WORLD_SIZE', '1'))
model = deepspeed.init_inference(model.eval(),
dtype="bfloat16",
replace_with_kernel_inject=False,
mp_size=world_size)
inputs = tokenizer('中国的首都是北京\n德国的首都是柏林\n孟加拉的首都是', return_tensors='pt')
inputs = inputs.to(model.device)
pred = model.generate(**inputs, max_new_tokens=32000)
print(tokenizer.decode(pred.cpu()[0], skip_special_tokens=True))
```
##
# 性能测评（Performance Evaluation）
### 长文本理解
我们使用LongBench的LSHT、LCC、Multifiled_QA_ZH数据集，对 Nanbeige-16B-Base-32k 模型进行了测评，相较具有长文本理解能力的同参数规模Base模型取得了不错的结果。
We evaluated the Nanbeige-16B-Base-32k model using LSHT, LCC, and Multifiled_QA_ZH from LongBench datasets. Compared to the Base model of the same parameter size with long-context comprehension capabilities, it achieved impressive results.
| | LSHT (分类) | LCC (代码) | Multifiled_QA_ZH (问答) |
|----------------------------|--------|-------|------------------|
| Chinese-Llama2-13B-16k | 31.0 | 9.6 | 36.4 |
| Qwen-14B-Dynamnic_ntk-Logn | 16.0 | 66.7 | 30.0 |
| Nanbeige-16B-Base-32k | 40.3 | 70.7 | 47.2 |
##
# 局限性（Limitations）
虽然我们在训练过程中非常注重模型的安全性，力求确保其输出符合伦理和法律要求的文本，但由于模型大小和概率生成范式的限制，无法完全避免产生各种不符合预期的输出情况。这些输出可能包含偏见、歧视等有害内容，请勿传播这些内容。我们不承担因传播不良信息而导致的任何后果。
While we place great emphasis on the safety of the model during the training process, striving to ensure that its outputs align with ethical and legal requirements, it may not completely avoid generating unexpected outputs due to the model's size and probabilistic nature. These outputs may include harmful content such as bias or discrimination. Please don't propagate such content. We do not assume any responsibility for the consequences resulting from the dissemination of inappropriate information.
# 协议（License）
使用 Nanbeige 模型时，您必须遵守 Apache 2.0 许可证和[《南北阁大语言模型许可协议》](https://huggingface.co/Nanbeige/Nanbeige-16B-Base-32k/resolve/main/南北阁大语言模型许可协议.pdf)。如果您打算将 Nanbeige 模型或其衍生产品用于商业目的，请通过以下联系邮箱 nanbeige@126.com 提交申请材料，以满足《南北阁大语言模型许可协议》的要求。经过审核后，我们将授予您非排他性、全球范围内、不可转让、不可再许可、可撤销的商业版权许可。
When using the Nanbeige models, you must comply with the Apache 2.0 License and the [License Agreement for Large Language Models Nanbeige](https://huggingface.co/Nanbeige/Nanbeige-16B-Base-32k/resolve/main/License_Agreement_for_Large_Language_Models_Nanbeige.pdf). If you intend to use the Nanbeige Models or its derivatives for commercial purposes, please submit application materials to meet the requirements of the Nanbeige Models Community License Agreement by contacting nanbeige@126.com. After review, We will grant you a non-exclusive, worldwide, non-transferable, non-sublicensable and revocable commercial copyright license.