language:
- en
- ko
license: other
tags:
- facebook
- meta
- pytorch
- llama
- llama-3
- llama-3-ko
- GGUF
pipeline_tag: text-generation
license_name: llama3
license_link: LICENSE
quantized_by: andrijdavid
Llama-3-Open-Ko-8B-Instruct-preview-GGUF
- Original model: Llama-3-Open-Ko-8B-Instruct-preview
Description
This repo contains GGUF format model files for Llama-3-Open-Ko-8B-Instruct-preview.
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. This is the source project for GGUF, providing both a Command Line Interface (CLI) and a server option.
- text-generation-webui, Known as the most widely used web UI, this project boasts numerous features and powerful extensions, and supports GPU acceleration.
- Ollama Ollama is a lightweight and extensible framework designed for building and running language models locally. It features a simple API for creating, managing, and executing models, along with a library of pre-built models for use in various applicationsβ
- KoboldCpp, A comprehensive web UI offering GPU acceleration across all platforms and architectures, particularly renowned for storytelling.
- GPT4All, This is a free and open source GUI that runs locally, supporting Windows, Linux, and macOS with full GPU acceleration.
- LM Studio An intuitive and powerful local GUI for Windows and macOS (Silicon), featuring GPU acceleration.
- LoLLMS Web UI. A notable web UI with a variety of unique features, including a comprehensive model library for easy model selection.
- Faraday.dev, An attractive, user-friendly character-based chat GUI for Windows and macOS (both Silicon and Intel), also offering GPU acceleration.
- llama-cpp-python, A Python library equipped with GPU acceleration, LangChain support, and an OpenAI-compatible API server.
- candle, A Rust-based ML framework focusing on performance, including GPU support, and designed for ease of use.
- ctransformers, A Python library featuring GPU acceleration, LangChain support, and an OpenAI-compatible AI server.
- localGPT An open-source initiative enabling private conversations with documents.
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.
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 folder.
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: LiteLLMs/Llama-3-Open-Ko-8B-Instruct-preview-GGUF and below it, a specific filename to download, such as: Q4_0/Q4_0-00001-of-00009.gguf.
Then click Download.
On the command line, including multiple files at once
I recommend using the huggingface-hub
Python library:
pip3 install huggingface-hub
Then you can download any individual model file to the current directory, at high speed, with a command like this:
huggingface-cli download LiteLLMs/Llama-3-Open-Ko-8B-Instruct-preview-GGUF Q4_0/Q4_0-00001-of-00009.gguf --local-dir . --local-dir-use-symlinks False
More advanced huggingface-cli download usage (click to read)
You can also download multiple files at once with a pattern:
huggingface-cli download LiteLLMs/Llama-3-Open-Ko-8B-Instruct-preview-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.
To accelerate downloads on fast connections (1Gbit/s or higher), install hf_transfer
:
pip3 install huggingface_hub[hf_transfer]
And set environment variable HF_HUB_ENABLE_HF_TRANSFER
to 1
:
HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli download LiteLLMs/Llama-3-Open-Ko-8B-Instruct-preview-GGUF Q4_0/Q4_0-00001-of-00009.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.
Make sure you are using llama.cpp
from commit d0cee0d or later.
./main -ngl 35 -m Q4_0/Q4_0-00001-of-00009.gguf --color -c 8192 --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 8192
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. Note that longer sequence lengths require much more resources, so you may need to reduce this value.
If you want to have a chat-style conversation, replace the -p <PROMPT>
argument with -i -ins
For other parameters and how to use them, please refer to the llama.cpp documentation
How to run in text-generation-webui
Further instructions can be found in the text-generation-webui documentation, here: text-generation-webui/docs/04 β Model Tab.md.
How to run from Python code
You can use GGUF models from Python using the llama-cpp-python or ctransformers libraries. Note that at the time of writing (Nov 27th 2023), ctransformers has not been updated for some time and is not compatible with some recent models. Therefore I recommend you use llama-cpp-python.
How to load this model in Python code, using llama-cpp-python
For full documentation, please see: llama-cpp-python docs.
First install the package
Run one of the following commands, according to your system:
# Base ctransformers with no GPU acceleration
pip install llama-cpp-python
# With NVidia CUDA acceleration
CMAKE_ARGS="-DLLAMA_CUBLAS=on" pip install llama-cpp-python
# Or with OpenBLAS acceleration
CMAKE_ARGS="-DLLAMA_BLAS=ON -DLLAMA_BLAS_VENDOR=OpenBLAS" pip install llama-cpp-python
# Or with CLBLast acceleration
CMAKE_ARGS="-DLLAMA_CLBLAST=on" pip install llama-cpp-python
# Or with AMD ROCm GPU acceleration (Linux only)
CMAKE_ARGS="-DLLAMA_HIPBLAS=on" pip install llama-cpp-python
# Or with Metal GPU acceleration for macOS systems only
CMAKE_ARGS="-DLLAMA_METAL=on" pip install llama-cpp-python
# In windows, to set the variables CMAKE_ARGS in PowerShell, follow this format; eg for NVidia CUDA:
$env:CMAKE_ARGS = "-DLLAMA_OPENBLAS=on"
pip install llama-cpp-python
Simple llama-cpp-python example code
from llama_cpp import Llama
# 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 = Llama(
model_path="./Q4_0/Q4_0-00001-of-00009.gguf", # Download the model file first
n_ctx=32768, # The max sequence length to use - note that longer sequence lengths require much more resources
n_threads=8, # The number of CPU threads to use, tailor to your system and the resulting performance
n_gpu_layers=35 # The number of layers to offload to GPU, if you have GPU acceleration available
)
# Simple inference example
output = llm(
"<PROMPT>", # Prompt
max_tokens=512, # Generate up to 512 tokens
stop=["</s>"], # Example stop token - not necessarily correct for this specific model! Please check before using.
echo=True # Whether to echo the prompt
)
# Chat Completion API
llm = Llama(model_path="./Q4_0/Q4_0-00001-of-00009.gguf", chat_format="llama-2") # Set chat_format according to the model you are using
llm.create_chat_completion(
messages = [
{"role": "system", "content": "You are a story writing assistant."},
{
"role": "user",
"content": "Write a story about llamas."
}
]
)
How to use with LangChain
Here are guides on using llama-cpp-python and ctransformers with LangChain:
Original model card: Llama-3-Open-Ko-8B-Instruct-preview
Llama-3-Open-Ko-8B-Instruct-preview
Update @ 2024.05.01: Pre-Release Llama-3-KoEn-8B model & Llama-3-KoEn-8B-Instruct-preview
Update @ 2024.04.24: Release Llama-3-Open-Ko-8B model & Llama-3-Open-Ko-8B-Instruct-preview
Model Details
Llama-3-Open-Ko-8B-Instruct-preview
Llama-3-Open-Ko-8B model is continued pretrained language model based on Llama-3-8B.
This model is trained fully with publicily available resource, with 60GB+ of deduplicated texts.
With the new Llama-3 tokenizer, the pretraining conducted with 17.7B+ tokens, which slightly more than Korean tokenizer(Llama-2-Ko tokenizer).
The train was done on TPUv5e-256, with the warm support from TRC program by Google.
With applying the idea from Chat Vector paper, I released Instruction model named Llama-3-Open-Ko-8B-Instruct-preview.
Since it is NOT finetuned with any Korean instruction set(indeed preview
), but it would be great starting point for creating new Chat/Instruct models.
Sample usage
from transformers import AutoTokenizer, AutoModelForCausalLM
import torch
model_id = "beomi/Llama-3-Open-Ko-8B-Instruct-preview"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
model_id,
torch_dtype="auto",
device_map="auto",
)
messages = [
{"role": "system", "content": "μΉμ ν μ±λ΄μΌλ‘μ μλλ°©μ μμ²μ μ΅λν μμΈνκ³ μΉμ νκ² λ΅νμ. λͺ¨λ λλ΅μ νκ΅μ΄(Korean)μΌλ‘ λλ΅ν΄μ€."},
{"role": "user", "content": "νΌλ³΄λμΉ μμ΄μ΄ λμΌ? κ·Έλ¦¬κ³ νΌλ³΄λμΉ μμ΄μ λν΄ νμ΄μ¬ μ½λλ₯Ό μ§μ€λ³Όλ?"},
]
input_ids = tokenizer.apply_chat_template(
messages,
add_generation_prompt=True,
return_tensors="pt"
).to(model.device)
terminators = [
tokenizer.eos_token_id,
tokenizer.convert_tokens_to_ids("<|eot_id|>")
]
outputs = model.generate(
input_ids,
max_new_tokens=512,
eos_token_id=terminators,
do_sample=True,
temperature=1,
top_p=0.9,
)
response = outputs[0][input_ids.shape[-1]:]
print(tokenizer.decode(response, skip_special_tokens=True))
Sample Output
μ¬λ¬λΆκ»μλ λ¬Όλ‘ μμλμ§λΌλ μ€λͺ
μ λλ¦¬κ² μ΅λλ€! νΌλ³΄λμΉ(Piconacci) μμ΄μ λ€μκ³Ό κ°μ μμ΄μ λ§ν©λλ€: 0, 1, 1, 2, 3, 5, 8, 13, 21, 34, 55, 89, 144, 233,.... μ΄λ κ² κ³μλκ³ μμ΅λλ€. μ΄ μμ΄μ μΌλ°μ μΌλ‘λ μνμμ 무νν μ§μλ©λλ€. κ° μλ μ΄μ μμ λ λ°°κ° λλ μμ΄μ
λλ€.
μ΄ μμ΄μ νμ΄μ¬(Python)μΌλ‘ ꡬννκ³ μ νμ λ€λ©΄ λ€μκ³Ό κ°μ μ½λλ₯Ό μμ±ν μ μμ΅λλ€.
```python
def fibonacci(n):
if n == 0:
return 0
elif n == 1:
return 1
else:
return fibonacci(n-1) + fibonacci(n-2)
```
μ΄ μ½λμμλ `n`μ΄ μ
λ ₯λ°μ μλ₯Ό λνλ΄λ λ³μλ‘, νΌλ³΄λμΉ μμ΄μ ꡬνν©λλ€. μ΄ μ½λλ₯Ό μ€νμν€λ©΄ μ
λ ₯λ°μ μμ λ°λΌ νΌλ³΄λμΉ μμ΄μ ν΄λΉ νμ λ°νν©λλ€. μλ₯Ό λ€μ΄, `fibonacci(10)` νλ©΄ 55λ₯Ό λ°νν©λλ€.
μ΄ μ½λλ μ¬κ·μ ν¨μμ΄κΈ° λλ¬Έμ λ©λͺ¨λ¦¬ μ¬μ©μ΄ μ€μ΄λ€ μ μμ΅λλ€. κ·Έλ¬λ μ¬κ·μ ν¨μλ λ©λͺ¨λ¦¬ μ¬μ©λμ΄ λ§μμ§ μ μμ΅λλ€. μ΄λ₯Ό μ€μ΄κΈ° μν΄ μ΄ν°λ μ΄ν°λΈ ν¨μλ‘ κ΅¬νν μλ μμ΅λλ€.
```python
a, b = 0, 1
for i in range(n):
a, b = b, a + b
```
μ΄ μ½λμμλ μ΄ν°λ μ΄ν°λΈ ν¨μλ‘ νΌλ³΄λμΉ μμ΄μ ꡬνν©λλ€. μ΄ μ½λλ₯Ό μ€νμν€λ©΄ μ
λ ₯λ°μ μμ λ°λΌ νΌλ³΄λμΉ μμ΄μ ν΄λΉ νμ λ°νν©λλ€. μ΄ μ½λλ λ©λͺ¨λ¦¬ μ¬μ©λμ μ€μ
λλ€. λ€λ§, μ΄ μ½λλ λ 볡μ‘ν©λλ€. μ΄ μ½λλ₯Ό λ κ°λ¨νκ² νλ €λ©΄ μ΄ν°λ μ΄ν°λΈ ν¨μλ₯Ό λ μ ꡬνν μ μμ΅λλ€.
```python
a, b = 0, 1
while True:
if n == 1:
return a
a, b = b, a + b
n -= 1
```
μ΄ μ½λμμλ μ
λ ₯λ°μ μμ λ°λΌ νΌλ³΄λμΉ μμ΄μ...
More samples
I used same system prompt, but you could change on your own.
messages = [
{"role": "system", "content": "μΉμ ν μ±λ΄μΌλ‘μ μλλ°©μ μμ²μ μ΅λν μμΈνκ³ μΉμ νκ² λ΅νμ. λͺ¨λ λλ΅μ νκ΅μ΄(Korean)μΌλ‘ λλ΅ν΄μ€."},
{"role": "user", "content": "νκ΅μ μ ν νλ²μ λν΄μ μμΈν μ€λͺ
ν΄μ£ΌμΈμ."},
]
μ΄μμ μ₯κ΅°μ μ‘°μ μλμ 무μΈμΌλ‘, λνλ―Όκ΅μ κ΅κ°μ 곡μμΈ μ°μμ 곡 μ΄μμ μ₯κ΅°μ 1545λ
9μ 28μΌ κ²½μλ¨λ μλ Ήμμ νμ΄λ¬μ΅λλ€. μ΄μμ μ₯κ΅°μ ν΅μ μ¬ μμ μ μλ λ μ‘°μ κ³Ό λͺ
λλΌ μ°ν©κ΅°μ λκ·λͺ¨ κ΅°λλ₯Ό μΌμΌμΌ λ첩μ μ±κ³΅μ μΌλ‘ μ΄λμ΄ μ μ μ¬λκ³Ό μμ§μλμ μΉλ¦¬λ‘ μ΄λ μΈλ¬Όμ
λλ€. κ·Έλ 1592λ
μ λΌμ’μμκ΄μ°°μ¬κ° λμ΄ μ λΌμ’μμμμ μμ μ 물리쳀μΌλ©°, 1597λ
μλ μμκ³Ό νμ λ°©μ΄μ μ±κ³΅μ μΌλ‘ μΉλ£νμ΅λλ€. λν λͺ
λλ첩μμ μμ κ³Ό 격μ νμ¬ μ΄μμ μ΄μμ μ₯κ΅°μ λͺ
λλ첩μμ μ‘°μ κ΄κ΅°μ μΉλ¦¬λ₯Ό μ΄λμμ΅λλ€. μ΄μμ μ₯κ΅°μ μ리λ₯Ό μ§ν€κΈ° μν΄ μ 리νμ κ³ μνλ κ²°λ¨μ λ΄λ Έμ΅λλ€. μ΄μμ μ₯κ΅°μ 1598λ
μ¬μ²μ± μ ν¬μμ ν¨μ ν ν μ μΈκ° μμ λΌ μ μΈκ° λΆλ¦¬ν΄μ§μ, λ¨ννμ¬ μ΄μμ μ₯κ΅°μ κ²°κ΅ μΆ©λ¬΄κ³΅ μ΄μμ μ μΉνΈλ₯Ό λ°μμ΅λλ€. κ·Έμ 곡μ μ λνλ―Όκ΅ μ΄μμ μ₯κ΅° κΈ°λ
κ΄μΌλ‘ λͺ
μλ₯Ό λμ΄κ³ μμ΅λλ€. μ΄μμ μ₯κ΅°λμ μ νμ μΉνΈλ‘ 1963λ
건κ΅νμ₯ μ΅κ³ νμ₯μ μμ¬λ°μμΌλ©°, λνλ―Όκ΅μ κ΅λ³΄ μ 13νΈλ‘ μ§μ λμ΅λλ€. μ΄μμ μ₯κ΅°μ μ μ€μ μΈ μΈλ¬Όλ‘ νκ΅μ μμ¬μμ ν° μ‘±μ μ λ¨κ²Όμ΅λλ€. μ΄μμ μ₯κ΅°μ΄ μλ λΉμμ μ ν¬μ₯μλ μΆ©λ¬΄κ³΅μ΄ μ λΌ μ²μμκ² λλΌλ₯Ό μν΄ μΈμ΄ κ³³μ μ΄μμ μ₯κ΅°μ λ¬κ° μμ΅λλ€. λλΌμ κ³ λμ μ§μ± μ΄μμ μ₯κ΅°λμ μμ
κ³Ό μ리λ₯Ό κΈ°λ
νλ κ³³μ
λλ€. μ΄μμ μ₯κ΅°μ νμ μ μ΄μμ μ₯κ΅° κΈ°λ
κ΄, μ΄μμ 곡μ λ±μ΄ μμ΅λλ€. μ΄μμ μ₯κ΅°μ 곡μ μ κΈ°λ
νκΈ° μν΄ λ€μν λ¬Ένμ μ°κ³Ό κΈ°λ
λ¬Όμ΄ μ‘°μ±λμ΅λλ€. μ΄μμ μ₯κ΅°μ μ μ μ λλΌλ₯Ό μν΄ μΈμ΄ μλν μ°λ¦¬μ μμ
μ
λλ€. 1552λ
12μμλ μ΄μμ μ₯κ΅°μ μ μ€μ μλ λλ€μ λ°°ν₯, νμ¬λ 좩무곡 μ΄μμ μ₯κ΅°κ»μλ μ°λ¦¬λλΌμ μμ§μ μΈλ¬Όμ΄ λμ΅λλ€. μ΄μμ μ₯κ΅°μ μ‘°μ μμ‘°μ μ€μλ₯Ό λ°κ³ μμ΅λλ€. μ΄μμ μ₯κ΅°μ μ μ‘°μκ² λ§μ λμλ₯Ό μ΄λ£¨μμ΅λλ€. μ΄μμ μ₯κ΅°μ 곡μ μ μ€λλ μ μ΄λμ΄ μ£Όλ μΈλ¬Όμ
λλ€. κ·Έλ₯Ό λͺ¨μ κ²μ μ κ΅μ μ¬λλ€μ
λλ€. μ΄μμ μ₯κ΅°λμ κ³ κ²°ν μ μ μ μ°λ¦¬λλΌκ° μ΄λ£©ν μ±κ³Όμ
λλ€. μ΄μμ μ₯κ΅°μ λλΌλ₯Ό μν΄ μΈμ΄ μλν μΈλ¬Όμ
λλ€. μ΄μμ μ₯κ΅°μ λνλ―Όκ΅μμ κ°μ₯ μλν μΈλ¬Όμ
λλ€. μ΄μμ μ₯κ΅°μ νμ μ μμν κΈ°μ΅ λκ² μ§μ.
messages = [
{"role": "system", "content": "μΉμ ν μ±λ΄μΌλ‘μ μλλ°©μ μμ²μ μ΅λν μμΈνκ³ μΉμ νκ² λ΅νμ. λͺ¨λ λλ΅μ νκ΅μ΄(Korean)μΌλ‘ λλ΅ν΄μ€."},
{"role": "user", "content": "μΈμ’
λμμ μ§νμ λ§₯λΆ λμ§ μ¬κ±΄μ λν΄μ μκ³ μμ΄?"},
]
μ¬κ³Όμ μλμ νμΈνκ² μ΅λλ€. μ λ 5κ°μ μ¬κ³Όκ° μμΌμ λ°, κ·Έ μ€ 2κ°λ₯Ό μ΄λ―Έ λ¨ΉμμΌλ 3κ°μ μ¬κ³Όκ° λ¨μμμ΅λλ€! γ
γ
. κ³μν΄μ λ¨Ήμ κ³ν μμμμ? π