metadata

thumbnail: >-
  https://assets-global.website-files.com/646b351987a8d8ce158d1940/64ec9e96b4334c0e1ac41504_Logo%20with%20white%20text.svg
metrics:
  - memory_disk
  - memory_inference
  - inference_latency
  - inference_throughput
  - inference_CO2_emissions
  - inference_energy_consumption
tags:
  - pruna-ai

This repo contains GGUF versions of the lightblue/suzume-llama-3-8B-japanese model.

Simply make AI models cheaper, smaller, faster, and greener!

Give a thumbs up if you like this model!
Contact us and tell us which model to compress next here.
Request access to easily compress your own AI models here.
Read the documentations to know more here
Join Pruna AI community on Discord here to share feedback/suggestions or get help.

Frequently Asked Questions

How does the compression work? The model is compressed with GGUF.
How does the model quality change? The quality of the model output might vary compared to the base model.
What is the model format? We use GGUF format.
What calibration data has been used? If needed by the compression method, we used WikiText as the calibration data.
How to compress my own models? You can request premium access to more compression methods and tech support for your specific use-cases here.

Downloading and running the models

You can download the individual files from the Files & versions section. Here is a list of the different versions we provide. For more info checkout this chart and this guide:

Quant type	Description
Q5_K_M	High quality, recommended.
Q5_K_S	High quality, recommended.
Q4_K_M	Good quality, uses about 4.83 bits per weight, recommended.
Q4_K_S	Slightly lower quality with more space savings, recommended.
IQ4_NL	Decent quality, slightly smaller than Q4_K_S with similar performance, recommended.
IQ4_XS	Decent quality, smaller than Q4_K_S with similar performance, recommended.
Q3_K_L	Lower quality but usable, good for low RAM availability.
Q3_K_M	Even lower quality.
IQ3_M	Medium-low quality, new method with decent performance comparable to Q3_K_M.
IQ3_S	Lower quality, new method with decent performance, recommended over Q3_K_S quant, same size with better performance.
Q3_K_S	Low quality, not recommended.
IQ3_XS	Lower quality, new method with decent performance, slightly better than Q3_K_S.
Q2_K	Very low quality but surprisingly usable.

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

Option A - Downloading in text-generation-webui:
Step 1: Under Download Model, you can enter the model repo: lightblue-suzume-llama-3-8B-japanese-GGUF-smashed and below it, a specific filename to download, such as: phi-2.IQ3_M.gguf.
Step 2: Then click Download.
Option B - Downloading on the command line (including multiple files at once):
Step 1: We recommend using the huggingface-hub Python library:

pip3 install huggingface-hub

Step 2: Then you can download any individual model file to the current directory, at high speed, with a command like this:

huggingface-cli download lightblue-suzume-llama-3-8B-japanese-GGUF-smashed suzume-llama-3-8B-japanese.IQ3_M.gguf --local-dir . --local-dir-use-symlinks False

More advanced huggingface-cli download usage (click to read)

Alternatively, you can also download multiple files at once with a pattern:

huggingface-cli download lightblue-suzume-llama-3-8B-japanese-GGUF-smashed --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 hf_transfer

And set environment variable HF_HUB_ENABLE_HF_TRANSFER to 1:

HF_HUB_ENABLE_HF_TRANSFER=1 huggingface-cli download lightblue-suzume-llama-3-8B-japanese-GGUF-smashed suzume-llama-3-8B-japanese.IQ3_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.

How to run model in GGUF format?

Option A - Introductory example with llama.cpp command

Make sure you are using llama.cpp from commit d0cee0d or later.

./main -ngl 35 -m suzume-llama-3-8B-japanese.IQ3_M.gguf --color -c 32768 --temp 0.7 --repeat_penalty 1.1 -n -1 -p "<s>[INST] {{prompt\}} [/INST]"

Change -ngl 32 to the number of layers to offload to GPU. Remove it if you don't have GPU acceleration.

Change -c 32768 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

Option B - Running in text-generation-webui

Further instructions can be found in the text-generation-webui documentation, here: text-generation-webui/docs/04 ‐ Model Tab.md.

Option C - Running 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](https://abetlen.github.io/llama-cpp-python/).

#### First install the package

Run one of the following commands, according to your system:

```shell
# 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

```python
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="./suzume-llama-3-8B-japanese.IQ3_M.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(
"<s>[INST] {{prompt}} [/INST]", # 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="./suzume-llama-3-8B-japanese.IQ3_M.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."
        }}
    ]
)
```

Option D - Running with LangChain

Here are guides on using llama-cpp-python and ctransformers with LangChain:

Configurations

The configuration info are in smash_config.json.

Credits & License

The license of the smashed model follows the license of the original model. Please check the license of the original model before using this model which provided the base model. The license of the pruna-engine is here on Pypi.

Want to compress other models?

Contact us and tell us which model to compress next here.
Request access to easily compress your own AI models here.