Instructions to use grapevine-AI/CALM3-22B-Chat-GGUF with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use grapevine-AI/CALM3-22B-Chat-GGUF with llama-cpp-python:

# !pip install llama-cpp-python

from llama_cpp import Llama

llm = Llama.from_pretrained(
	repo_id="grapevine-AI/CALM3-22B-Chat-GGUF",
	filename="calm3-22b-chat-Q4_K_M.gguf",
)

llm.create_chat_completion(
	messages = "No input example has been defined for this model task."
)

Notebooks
Google Colab
Kaggle
Local Apps

llama.cpp

How to use grapevine-AI/CALM3-22B-Chat-GGUF with llama.cpp:

Install from brew

brew install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf grapevine-AI/CALM3-22B-Chat-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf grapevine-AI/CALM3-22B-Chat-GGUF:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama-server -hf grapevine-AI/CALM3-22B-Chat-GGUF:Q4_K_M
# Run inference directly in the terminal:
llama-cli -hf grapevine-AI/CALM3-22B-Chat-GGUF:Q4_K_M

Use pre-built binary

# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases
# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf grapevine-AI/CALM3-22B-Chat-GGUF:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf grapevine-AI/CALM3-22B-Chat-GGUF:Q4_K_M

Build from source code

git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli
# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf grapevine-AI/CALM3-22B-Chat-GGUF:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf grapevine-AI/CALM3-22B-Chat-GGUF:Q4_K_M

Use Docker

docker model run hf.co/grapevine-AI/CALM3-22B-Chat-GGUF:Q4_K_M

LM Studio
Jan
Ollama
How to use grapevine-AI/CALM3-22B-Chat-GGUF with Ollama:
```
ollama run hf.co/grapevine-AI/CALM3-22B-Chat-GGUF:Q4_K_M
```

Unsloth Studio new

How to use grapevine-AI/CALM3-22B-Chat-GGUF with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for grapevine-AI/CALM3-22B-Chat-GGUF to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for grapevine-AI/CALM3-22B-Chat-GGUF to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for grapevine-AI/CALM3-22B-Chat-GGUF to start chatting

Docker Model Runner
How to use grapevine-AI/CALM3-22B-Chat-GGUF with Docker Model Runner:
```
docker model run hf.co/grapevine-AI/CALM3-22B-Chat-GGUF:Q4_K_M
```

Lemonade

How to use grapevine-AI/CALM3-22B-Chat-GGUF with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull grapevine-AI/CALM3-22B-Chat-GGUF:Q4_K_M

Run and chat with the model

lemonade run user.CALM3-22B-Chat-GGUF-Q4_K_M

List all available models

lemonade list

caution!

このGGUFは本来の性能を十分に発揮できていない「暫定版」です。
これは2024年7月3日現在のllama.cppがCALM3モデル固有のpre-tokenization（≒前処理）をサポートしていないことに起因します。
妥協策として、pre-tokenization処理はllama.cppデフォルトのものを利用するように改造してありますが、これはモデルの性能低下を引き落としている可能性が極めて高いです。

What is this?

CyberAgentの日英バイリンガル言語モデルCyberAgentLM3-22B-ChatをGGUFフォーマットに変換したものです。

imatrix dataset

日本語能力を重視し、日本語が多量に含まれるTFMC/imatrix-dataset-for-japanese-llmデータセットを使用しました。
また、imatrixの算出においてはq8_0精度のモデルを使用しました。これは、現行のCUDA版llama.cppがbf16での計算に非対応であり、またf16を使用した場合も原因不明のnanが検出されたためです。

Chat template

<|im_start|>system
ここにSystem Promptを書きます。<|im_end|>
<|im_start|>user
ここにMessageを書きます。<|im_end|>
<|im_start|>assistant

Environment

Windows版llama.cpp-b3274およびpre-tokenization関連に細工を施した改造版convert-hf-to-gguf.pyを使用して量子化作業を実施しました。

License

apache-2.0

Developer

CyberAgent

Downloads last month: 41

GGUF

Model size

23B params

Architecture

llama

Hardware compatibility

4-bit

5-bit

6-bit

16-bit

Inference Providers NEW

This model isn't deployed by any Inference Provider. 🙋 Ask for provider support