Instructions to use grimjim/kukulemon-spiked-9B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use grimjim/kukulemon-spiked-9B with Transformers:

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

pipe = pipeline("text-generation", model="grimjim/kukulemon-spiked-9B")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("grimjim/kukulemon-spiked-9B")
model = AutoModelForCausalLM.from_pretrained("grimjim/kukulemon-spiked-9B")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use grimjim/kukulemon-spiked-9B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "grimjim/kukulemon-spiked-9B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "grimjim/kukulemon-spiked-9B",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/grimjim/kukulemon-spiked-9B

SGLang

How to use grimjim/kukulemon-spiked-9B 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 "grimjim/kukulemon-spiked-9B" \
    --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": "grimjim/kukulemon-spiked-9B",
		"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 "grimjim/kukulemon-spiked-9B" \
        --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": "grimjim/kukulemon-spiked-9B",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use grimjim/kukulemon-spiked-9B with Docker Model Runner:
```
docker model run hf.co/grimjim/kukulemon-spiked-9B
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

kululemon-spiked-9B

This is a frankenmerge of a pre-trained language model created using mergekit. As an experiment, this appears to be a partial success.

Lightly tested with temperature 1-1.2 and minP 0.01 with ChatML prompts; the model supports Alpaca prompts and has 8K context length, a result of its Mistral v0.1 provenance. The model's output has been coherent and stable during aforementioned testing.

The merge formula for this frankenmerge is below. It is conjectured that the shorter first section is not key to variation, the middle segment is key to balancing reasoning and variation, and that the lengthy final section is required for convergence and eventual stability. The internal instability is probably better suited for narrative involving unstable and/or unhinged characters and situations.

Quants available:

Merge Details

Merge Method

This model was merged using the passthrough merge method.

Models Merged

The following models were included in the merge:

grimjim/kukulemon-7B

Configuration

The following YAML configuration was used to produce this model:

slices:
  - sources:
    - model: grimjim/kukulemon-7B
      layer_range: [0, 12]
  - sources:
    - model: grimjim/kukulemon-7B
      layer_range: [8, 16]
  - sources:
    - model: grimjim/kukulemon-7B
      layer_range: [12, 32]
merge_method: passthrough
dtype: float16