Instructions to use mdomina/Kalithos-C1-SFT with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use mdomina/Kalithos-C1-SFT with Transformers:

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

pipe = pipeline("text-generation", model="mdomina/Kalithos-C1-SFT")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("mdomina/Kalithos-C1-SFT")
model = AutoModelForCausalLM.from_pretrained("mdomina/Kalithos-C1-SFT", 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=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use mdomina/Kalithos-C1-SFT with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "mdomina/Kalithos-C1-SFT"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "mdomina/Kalithos-C1-SFT",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/mdomina/Kalithos-C1-SFT

SGLang

How to use mdomina/Kalithos-C1-SFT 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 "mdomina/Kalithos-C1-SFT" \
    --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": "mdomina/Kalithos-C1-SFT",
		"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 "mdomina/Kalithos-C1-SFT" \
        --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": "mdomina/Kalithos-C1-SFT",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use mdomina/Kalithos-C1-SFT with Docker Model Runner:
```
docker model run hf.co/mdomina/Kalithos-C1-SFT
```

Qwen3-Coder-Next-Cyber-SFT

mdomina/Kalithos-C1 con l'adapter comportamentale SFT-LoRA fuso nei pesi (red/blue-team, ragionamento <think>, verdetto / MITRE / azione).

Modello self-contained = base cyber + comportamento SFT. È il punto di partenza per il GRPO/RLVR (la KL-reference). Equivalente a caricare base + Qwen3-Coder-Next-Cyber-SFT-lora, ma già fuso.

Merge: 48 moduli attention (q/k/v/o_proj), scaling α/r = 2.0.
Inferenza: usare repetition_penalty ≈ 1.15 e template qwen3_coder.
Benchmark (dal base+adapter): CyberMetric-500 92.8%.

Vedi mdomina/Kalithos-C1 (base) e il repo kalithos-cybersec (recipes/sft-combined/, recipes/grpo/).

Downloads last month: 16

Safetensors

Model size

85B params

Tensor type

BF16

Model tree for mdomina/Kalithos-C1-SFT

Base model

mdomina/Kalithos-C1

Finetuned

(1)

this model