Instructions to use Sethblocks/Starlight-Flash with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Sethblocks/Starlight-Flash with Transformers:

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

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

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Sethblocks/Starlight-Flash")
model = AutoModelForCausalLM.from_pretrained("Sethblocks/Starlight-Flash", device_map="auto")
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=256)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Sethblocks/Starlight-Flash with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/Sethblocks/Starlight-Flash

SGLang

How to use Sethblocks/Starlight-Flash 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 "Sethblocks/Starlight-Flash" \
    --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": "Sethblocks/Starlight-Flash",
		"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 "Sethblocks/Starlight-Flash" \
        --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": "Sethblocks/Starlight-Flash",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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

Starlight-Flash

My newer LLM based on Qwen-3.5 9B. It mostly focuses on personality and some general improvements. The model is not fully Uncensored but is less censored than Qwen.

Strengths

Much better than my last models at agentic tasks
Performs extremely fast as a very small model with relatively short CoT.
Much better than Qwen at following a Persona.

Weaknesses

Programming could be better, specifically for GUI/webui development.
Tends to require guidance, and sometimes gets confused.

Training

This model has taken the longest of all my models to develop, simply because a significant amount of the training code was rewritten, as well as creating an entirely new RL setup for this model.

GGUF: Sethblocks/Starlight-Flash-GGUF

Downloads last month: 357

Safetensors

Model size

9B params

Tensor type

BF16

Model tree for Sethblocks/Starlight-Flash

Quantizations

1 model