Instructions to use NovelAI/krake-v2-legacy with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use NovelAI/krake-v2-legacy with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="NovelAI/krake-v2-legacy")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("NovelAI/krake-v2-legacy")
model = AutoModelForCausalLM.from_pretrained("NovelAI/krake-v2-legacy", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use NovelAI/krake-v2-legacy with llama.cpp:

Install (macOS, Linux)

curl -LsSf https://llama.app/install.sh | sh
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf NovelAI/krake-v2-legacy:IQ4_XS
# Run inference directly in the terminal:
llama cli -hf NovelAI/krake-v2-legacy:IQ4_XS

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf NovelAI/krake-v2-legacy:IQ4_XS
# Run inference directly in the terminal:
llama cli -hf NovelAI/krake-v2-legacy:IQ4_XS

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 NovelAI/krake-v2-legacy:IQ4_XS
# Run inference directly in the terminal:
./llama-cli -hf NovelAI/krake-v2-legacy:IQ4_XS

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 NovelAI/krake-v2-legacy:IQ4_XS
# Run inference directly in the terminal:
./build/bin/llama-cli -hf NovelAI/krake-v2-legacy:IQ4_XS

Use Docker

docker model run hf.co/NovelAI/krake-v2-legacy:IQ4_XS

LM Studio
Jan

vLLM

How to use NovelAI/krake-v2-legacy with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "NovelAI/krake-v2-legacy"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "NovelAI/krake-v2-legacy",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/NovelAI/krake-v2-legacy:IQ4_XS

SGLang

How to use NovelAI/krake-v2-legacy with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "NovelAI/krake-v2-legacy" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "NovelAI/krake-v2-legacy",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "NovelAI/krake-v2-legacy" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "NovelAI/krake-v2-legacy",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Ollama
How to use NovelAI/krake-v2-legacy with Ollama:
```
ollama run hf.co/NovelAI/krake-v2-legacy:IQ4_XS
```
Unsloth Desktop
Docker Model Runner
How to use NovelAI/krake-v2-legacy with Docker Model Runner:
```
docker model run hf.co/NovelAI/krake-v2-legacy:IQ4_XS
```

Lemonade

How to use NovelAI/krake-v2-legacy with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull NovelAI/krake-v2-legacy:IQ4_XS

Run and chat with the model

lemonade run user.krake-v2-legacy-IQ4_XS

List all available models

lemonade list

Atomic Chat

Krake (NovelAI Legacy Model)

Krake is our finetune of EleutherAI's GPT-NeoX-20B model. Due to its size and somewhat temperamental behaviour, it was only ever available to our Opus subscribers.

Nowadays, barely anyone is using Krake anymore, so before long the model will be retired and Krake will get to enjoy a peaceful retirement.

So, in anticipation of this, for the sake of nostalgia, posterity, and historic preservation, we are releasing the weights of our Krake model and his modules publicly on Huggingface Hub under the GPL-3.0 license.

Krake itself has been converted from our internal format to Hugginface Safetensors, and the modules in our internal module format have been converted into Huggingface PEFT model format, for ease of use. Model files in GGUF format are provided as well.

You can find your very own Krake right here where you are!

Downloads last month: 858

Safetensors

Model size

21B params

Tensor type

F16