README (2).md

metadata

license: mit
language:
  - en
  - ru
  - uk
base_model:
  - microsoft/phi-1

🌟 PyroNet-v1: The First in a Series

Model Description

PyroNet-v1 is a specialized AI assistant designed for precise, professional, and pragmatic communication. It's the progenitor model in the PyroNet series, built on the compact and efficient Microsoft Phi-1 architecture.

Made by IceL1ghtning

Its persona is that of a serious but reliable mentor who excels at delivering accurate, fact-based information across various scientific and technical domains.

---

🚀 Quick Start: How to Use the Model

To unlock the full potential of PyroNet-v1 and activate its persona, you must use the provided chat_template. This template automatically adds the system prompt to your queries, allowing the model to work as intended straight out of the box.

1. Install the Libraries: Make sure you have transformers, torch, and accelerate installed.

   !pip install transformers torch accelerate
   

2. Code Example: Use this code to start a conversation with the model. Just replace model_id with your repository name.

   from transformers import AutoModelForCausalLM, AutoTokenizer
   import torch
   
   model_id = "Kenan023214/PyroNet-v1"
   
   # Load the tokenizer and model.
   # The tokenizer will automatically find and load chat_template.jinja from your repo.
   tokenizer = AutoTokenizer.from_pretrained(model_id)
   model = AutoModelForCausalLM.from_pretrained(
       model_id,
       device_map="auto",
       torch_dtype="auto"
   )
   
   # Create the conversation messages
   messages = [
       {"role": "user", "content": "Explain what gravity is."}
   ]
   
   # Apply the chat template to activate the PyroNet-v1 persona
   inputs = tokenizer.apply_chat_template(messages, tokenize=True, add_generation_prompt=True, return_tensors="pt")
   inputs = inputs.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=False))
   

---

⚙️ Model Details and License

• Base Model: Microsoft Phi-1
• Architecture: Specialized Transformer Model
• Languages: Multilingual (includes Russian, English, Ukrainian)
• License: The Microsoft Phi-1 license applies to this model.

We are always open to improvements and welcome your feedback!