nitinkore/phi-2-fine-tuned-product-decsription

Model Card for Model ID

How to use this model : Sample Code => Starts

import torch

from peft import PeftModel

from transformers import AutoModelForCausalLM, AutoTokenizer

Set device (CUDA if available)

device = torch.device("cuda" if torch.cuda.is_available() else "cpu")

Load tokenizer and base model

base_model = AutoModelForCausalLM.from_pretrained("microsoft/phi-2") model = PeftModel.from_pretrained(base_model, "nitinkore/phi-2-fine-tuned-product-decsription")

tokenizer = AutoTokenizer.from_pretrained("microsoft/phi-2")

Move model to CUDA

model.to(device)

Function to generate product description

def generate_product_description(category, max_length=256):

# Define prompt

eval_prompt = f"""
Given the product category, you need to generate a 'Product Description'.
### Category: {category}
### Product Description:
"""

# Tokenize and move input to device

model_input = tokenizer(eval_prompt, return_tensors="pt").to(device)

# Model in evaluation mode

model.eval()

# Generate with torch.no_grad()

with torch.no_grad():
    output = model.generate(
        **model_input,
        max_new_tokens=max_length,
        repetition_penalty=1.15,
        pad_token_id=tokenizer.eos_token_id  # Set pad_token_id to eos_token_id
    )

# Decode and return result

result = tokenizer.decode(output[0], skip_special_tokens=True)

# Extract only the generated description

result = result.split("### Product Description:")[1].strip()
return result

Example categories

categories = [ "Smartphones", "Laptops", "Headphones" ]

Generate descriptions

for category in categories:

description = generate_product_description(category)
print(f"Category: {category}\nDescription: {description}\n")

How to use this model : Sample Code => Ends

Model Card Authors [optional]

Nitin Kore

Model Card Contact

[email protected]

Framework versions

• PEFT 0.15.0