Gemma-model-fine-tuned-for-advertisement-generation

Gemma Fine-tuned for advertisement generation

This is a fine-tuned language model specifically trained for generating creative advertisement scripts and compelling call-to-actions. The model is part of the PromoGen project - an AI-powered advertisement generator.

Model Description

• Model Type: Fine-tuned Transformer Language Model
• Language: English
• Use Case: Advertisement script generation, marketing copy, call-to-action creation
• Base Model: Google's Gemma for CasualLLM
• Fine-tuning Dataset: Custom advertisement and marketing copy dataset

Model Performance

• Model Size: ~4.8GB
• Generation Time: 30-60 seconds (GPU) / 2-5 minutes (CPU)
• Memory Usage: ~6GB during inference
• Recommended Hardware: NVIDIA GPU with 8GB+ VRAM

Usage

Quick Start

from transformers import AutoTokenizer, AutoModelForCausalLM
import torch

# Load model and tokenizer
tokenizer = AutoTokenizer.from_pretrained("Bhavin1905/Promogen-finetuned-model")
model = AutoModelForCausalLM.from_pretrained("Bhavin1905/Promogen-finetuned-model")

# Set device
device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
model.to(device)

# Generate advertisement
def generate_ad(company_name, product_name, description):
    prompt = f"Company: {company_name}\nProduct: {product_name}\nDescription: {description}\nAdvertisement:"
    
    inputs = tokenizer.encode(prompt, return_tensors="pt").to(device)
    
    with torch.no_grad():
        outputs = model.generate(
            inputs,
            max_new_tokens=100,
            temperature=0.8,
            top_p=0.85,
            do_sample=True,
            pad_token_id=tokenizer.eos_token_id
        )
    
    generated_text = tokenizer.decode(outputs[0], skip_special_tokens=True)
    return generated_text.split("Advertisement:")[-1].strip()

# Example usage
ad_script = generate_ad(
    company_name="TechCorp",
    product_name="SmartWatch Pro",
    description="Advanced fitness tracking with heart rate monitoring"
)
print(ad_script)

Integration with PromoGen Application

This model is designed to work with the PromoGen Flask web application. To use it:

1. Clone the PromoGen repository:

   git clone https://github.com/cptghostriley/PromoGen.git
   cd PromoGen
   

2. Download this model:

   from huggingface_hub import snapshot_download
   snapshot_download(
       repo_id='Bhavin1905/Promogen-finetuned-model',
       local_dir='./promogen_final_model'
   )
   

3. Run the application:

   pip install -r requirements.txt
   python app.py
   

Model Details

Training Configuration

• Training Method: Fine-tuning on advertisement dataset
• Optimization: AdamW optimizer
• Learning Rate: [Training details]
• Batch Size: [Training details]
• Training Steps: [Training details]

Model Files

• config.json: Model configuration
• generation_config.json: Generation parameters
• model-00001-of-00002.safetensors: Model weights (part 1)
• model-00002-of-00002.safetensors: Model weights (part 2)
• model.safetensors.index.json: Model index
• special_tokens_map.json: Special tokens mapping
• tokenizer.json: Tokenizer configuration
• tokenizer.model: Tokenizer model file
• tokenizer_config.json: Tokenizer configuration

Hardware Requirements

Minimum Requirements

• RAM: 8GB+ (16GB recommended)
• Storage: 10GB+ free space
• CPU: Multi-core processor

Recommended for Best Performance

• GPU: NVIDIA GPU with 8GB+ VRAM (RTX 3070/4060 or better)
• RAM: 16GB+
• Storage: SSD with 20GB+ free space

Limitations and Bias

• The model is trained specifically for advertisement generation
• Performance may vary for other text generation tasks
• Generated content should be reviewed for appropriateness
• May reflect biases present in training data

Citation

If you use this model in your research or applications, please cite:

@misc{promogen2024,
  title={PromoGen: AI-Powered Advertisement Generator},
  author={Ashwin Mahawar, Bhavin, and Team PromoGen},
  year={2024},
  url={https://github.com/cptghostriley/PromoGen}
}

Contact

For questions, issues, or collaboration:

• GitHub Repository: PromoGen
• Issues: GitHub Issues

---

Note: This model is part of the PromoGen project and is designed for educational and research purposes. Please ensure compliance with applicable laws and regulations when using for commercial purposes.