Instructions to use Davide531/KyroLM-V2-Experimental with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Davide531/KyroLM-V2-Experimental with Transformers:

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

pipe = pipeline("text-generation", model="Davide531/KyroLM-V2-Experimental")

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("Davide531/KyroLM-V2-Experimental", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

llama.cpp

How to use Davide531/KyroLM-V2-Experimental 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 Davide531/KyroLM-V2-Experimental:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf Davide531/KyroLM-V2-Experimental:Q4_K_M

Install from WinGet (Windows)

winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf Davide531/KyroLM-V2-Experimental:Q4_K_M
# Run inference directly in the terminal:
llama cli -hf Davide531/KyroLM-V2-Experimental:Q4_K_M

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 Davide531/KyroLM-V2-Experimental:Q4_K_M
# Run inference directly in the terminal:
./llama-cli -hf Davide531/KyroLM-V2-Experimental:Q4_K_M

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 Davide531/KyroLM-V2-Experimental:Q4_K_M
# Run inference directly in the terminal:
./build/bin/llama-cli -hf Davide531/KyroLM-V2-Experimental:Q4_K_M

Use Docker

docker model run hf.co/Davide531/KyroLM-V2-Experimental:Q4_K_M

LM Studio
Jan

vLLM

How to use Davide531/KyroLM-V2-Experimental with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Davide531/KyroLM-V2-Experimental"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Davide531/KyroLM-V2-Experimental",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Davide531/KyroLM-V2-Experimental:Q4_K_M

SGLang

How to use Davide531/KyroLM-V2-Experimental 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 "Davide531/KyroLM-V2-Experimental" \
    --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": "Davide531/KyroLM-V2-Experimental",
		"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 "Davide531/KyroLM-V2-Experimental" \
        --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": "Davide531/KyroLM-V2-Experimental",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Ollama
How to use Davide531/KyroLM-V2-Experimental with Ollama:
```
ollama run hf.co/Davide531/KyroLM-V2-Experimental:Q4_K_M
```
Unsloth Desktop
Docker Model Runner
How to use Davide531/KyroLM-V2-Experimental with Docker Model Runner:
```
docker model run hf.co/Davide531/KyroLM-V2-Experimental:Q4_K_M
```

Lemonade

How to use Davide531/KyroLM-V2-Experimental with Lemonade:

Pull the model

# Download Lemonade from https://lemonade-server.ai/
lemonade pull Davide531/KyroLM-V2-Experimental:Q4_K_M

Run and chat with the model

lemonade run user.KyroLM-V2-Experimental-Q4_K_M

List all available models

lemonade list

Atomic Chat

KyroLM V2 Experimental

Model Details


Model	KyroLM V2 Experimental
Foundation	Qwen3-4B-Instruct-2507
Training data	KyroLM second-generation dataset
Type	General-purpose instruction-tuned LLM
Languages	English (primary); French, Italian, Spanish (experimental)
License	MIT

Description

KyroLM V2 Experimental is a general-purpose model made for everything: it is the first stable model of the V2 line and is intended to serve as the foundation for the upcoming Gen 2 models. It prioritizes being a solid, versatile base over being a "really good" standalone model. This AI is made to speak formarly , reduce hallucinations and make reasoning logic better .

Capabilities

Good at

Conversation
Basic knowledge
Very basic coding
Moderate chain-of-thought (CoT) and math
Logic

Support

Feature	Status
Reasoning	Supported
Tool calling	Partially supported

Strengths and Limitations

Pros

Good at logic
Moderate reasoning
Versatile: handles a wide range of everyday tasks

Cons

Low general knowledge
Reasoning and logic are only moderate on harder problems
Designed as a base for next-generation models, not as a final product
French, Italian and Spanish are experimental, so expect lower quality than in English
Tool calling is only partial and may be unreliable

Intended Use

General chat and assistant tasks
Light coding help and simple math
A foundation for further fine-tuning and for the next KyroLM Gen 2 models
Experimentation and testing

Not recommended for factual lookups that require broad or up-to-date knowledge, or for high-stakes decisions. Always verify important outputs.

Usage

from transformers import AutoModelForCausalLM, AutoTokenizer

model_id = "Davide531/KyroLM-V2-Experimental"

tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(
    model_id, torch_dtype="auto", device_map="auto"
)

messages = [
    {"role": "user", "content": "Explain step by step why 17 is a prime number."}
]
text = tokenizer.apply_chat_template(
    messages, tokenize=False, add_generation_prompt=True
)
inputs = tokenizer([text], return_tensors="pt").to(model.device)

output = model.generate(**inputs, max_new_tokens=512)
print(tokenizer.decode(output[0][inputs.input_ids.shape[1]:], skip_special_tokens=True))

Status

This is an experimental release. Behavior, quality and format may change in later KyroLM V2 series models.

Acknowledgements

Built on top of Qwen3-4B-Instruct-2507 by the Qwen team.

Downloads last month: 22

GGUF

Model size

4B params

Architecture

qwen3

Hardware compatibility

4-bit

Model tree for Davide531/KyroLM-V2-Experimental

Base model

Qwen/Qwen3-4B-Instruct-2507

Quantized

(334)

this model