README.md

metadata

license: mit
language:
  - ru
  - en
  - uk

🔥 PyroNet-mini

PyroNet-mini is a fine-tuned lightweight variant of microsoft/Phi-4-mini-instruct,
trained with a custom multilingual instruction dataset (English, Ukrainian, Russian).
It belongs to the PyroNet family, serving as the mini version of the main PyroNet.

---

✨ Features

• 🔹 Based on Phi-4-mini-instruct
• 🔹 Fine-tuned with LoRA adapters
• 🔹 Multilingual (English / Українська / Русский)
• 🔹 Softer safety alignment for more natural answers
• 🔹 Optimized for dialogue, coding help, and educational purposes

---

🚀 Usage

Inference with Transformers

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

model_id = "Kenan023214/PyroNet-mini"

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

messages = [
    {"role": "system", "content": system_prompt},
    {"role": "user", "content": "Explain quantum entanglement in simple words."}
]


input_text = tokenizer.apply_chat_template(messages, tokenize=False, add_generation_prompt=True)
inputs = tokenizer(input_text, return_tensors="pt").to(model.device)

# Generate the response
outputs = model.generate(
    inputs,
    max_new_tokens=256,
    pad_token_id=tokenizer.eos_token_id
)

print(tokenizer.decode(outputs[0], skip_special_tokens=True))

---

📊 Intended Use

Educational Q&A

Coding assistance

Multilingual chatbots

Creative writing and storytelling

---

⚡ Limitations

Being a mini model, it is less capable than larger PyroNet variants

May produce factual inaccuracies

Softer censorship may allow sensitive content

---

👤 Author & Contact

Author: IceL1ghtning

Maintainer: KenanAI

Contact:

---