README.md

metadata

language:
  - en
license: mit
tags:
  - text-generation
  - mistral
  - business
  - technology
  - assistant
  - conversational
  - safetensors
  - AI
  - problem-solving
pipeline_tag: text-generation
inference: true
widget:
  - text: How can I improve my business strategy?

SireIQ - Professional AI Assistant

🌟 About SireIQ

SireIQ is a professional AI assistant designed specifically for business, technology, and problem-solving tasks. Built on Mistral-7B-Instruct-v0.3 and fine-tuned on over 25,000 high-quality instruction samples, SireIQ provides clear, accurate, and actionable advice tailored for professionals and teams.

✨ Key Features

• Business Strategy: Help with go-to-market strategies, business planning, and decision-making
• Technology Guidance: Explain complex technical concepts, coding assistance, and architecture recommendations
• Problem-Solving: Step-by-step solutions for professional challenges
• Conversational: Natural, professional dialogue with proper system prompt support
• Lightweight: 7B parameters, optimized for fast inference

🚀 Quick Start

Using Transformers

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

# Load model and tokenizer
model_name = "tarvico/sireiq"
tokenizer = AutoTokenizer.from_pretrained(model_name)
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    torch_dtype=torch.float16,
    device_map="auto"
)

# Define system prompt and user query
system_prompt = "You are SireIQ, a professional AI assistant for business, technology, and problem-solving. You provide clear, accurate, and actionable advice tailored for professionals and teams."
user_query = "What are the key steps to build a go-to-market strategy for a SaaS product?"

# Format the conversation
messages = [
    {"role": "system", "content": system_prompt},
    {"role": "user", "content": user_query}
]
inputs = tokenizer.apply_chat_template(messages, return_tensors="pt").to(model.device)

# Generate response
outputs = model.generate(
    inputs,
    max_new_tokens=512,
    temperature=0.7,
    top_p=0.9,
    do_sample=True,
    pad_token_id=tokenizer.eos_token_id
)

# Print the result
response = tokenizer.decode(outputs[0], skip_special_tokens=True)
print(response)

Simple Inference

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch

tokenizer = AutoTokenizer.from_pretrained("tarvico/sireiq")
model = AutoModelForCausalLM.from_pretrained(
    "tarvico/sireiq",
    torch_dtype=torch.float16,
    device_map="auto"
)

prompt = "Explain AI in simple terms."
input_ids = tokenizer(prompt, return_tensors="pt").input_ids
output_ids = model.generate(input_ids, max_new_tokens=200)
print(tokenizer.decode(output_ids[0], skip_special_tokens=True))

📋 Model Details

Attribute	Value
Base Model
Parameters	7B
Fine-tuning Method	LoRA
Training Samples	25,000+
License	MIT
Language	English

📊 Training Data

SireIQ was fine-tuned on a diverse dataset including:

• Databricks Dolly-15k
• OpenHermes-2.5 (business/tech filtered)
• WizardLM Evol-Instruct-70k
• Alpaca-cleaned
• GSM8K (math reasoning)

🔒 License

This model is released under the MIT License. See the LICENSE file for more details.

📞 Support

For questions or issues, please open a discussion on the Hugging Face repo.

🤝 Contributing

Contributions to improve SireIQ are welcome! Please feel free to submit issues or pull requests.