Instructions to use Epiculous/NovaSpark with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Epiculous/NovaSpark with Transformers:

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

pipe = pipeline("text-generation", model="Epiculous/NovaSpark")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Epiculous/NovaSpark")
model = AutoModelForCausalLM.from_pretrained("Epiculous/NovaSpark")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Inference
Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use Epiculous/NovaSpark with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Epiculous/NovaSpark"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Epiculous/NovaSpark",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Epiculous/NovaSpark

SGLang

How to use Epiculous/NovaSpark 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 "Epiculous/NovaSpark" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Epiculous/NovaSpark",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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 "Epiculous/NovaSpark" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Epiculous/NovaSpark",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Epiculous/NovaSpark with Docker Model Runner:
```
docker model run hf.co/Epiculous/NovaSpark
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

Switching things up a bit since the last slew of models were all 12B, we now have NovaSpark! NovaSpark is an 8B model trained on GrimJim's abliterated version of arcee's SuperNova-lite. The hope is abliteration will remove some of the inherant refusals and censorship of the original model, however I noticed that finetuning on GrimJim's model undid some of the abliteration, therefore more than likely abiliteration will have to be reapplied to the resulting model to reinforce it.

Quants!

full / exl2 / gguf

Prompting

This model is trained on llama instruct template, the prompting structure goes a little something like this:

<|begin_of_text|><|start_header_id|>system<|end_header_id|>

{system_prompt}<|eot_id|><|start_header_id|>user<|end_header_id|>

{prompt}<|eot_id|><|start_header_id|>assistant<|end_header_id|>