The model has been finetuned to optimize the conversion of short prompts into detailed prompts, enabling stable diffusion or flux-based image generation. This refinement enables users to craft more specific and nuanced requests, resulting in higher-quality and more coherent images.

Model Description

Developed by: [Imran Ali]
Model type: [T5 (Text-to-Text Transfer Transformer)]
Language(s) (NLP): [English]
License: [apache-2.0]
Finetuned from model: [t5-small]
Demo: Demo Space

How to Get Started with the Model

Use the code below to get started with the model.

python

from transformers import T5Tokenizer, T5ForConditionalGeneration

# Load the tokenizer and model
tokenizer = T5Tokenizer.from_pretrained("imranali291/flux-prompt-enhancer")
model = T5ForConditionalGeneration.from_pretrained("imranali291/flux-prompt-enhancer")

# Example input
input_text = "Futuristic cityscape at twilight descent."

# Tokenize input
input_ids = tokenizer(input_text, return_tensors="pt").input_ids

# Generate output
output = model.generate(input_ids, max_length=128, eos_token_id=tokenizer.eos_token_id, do_sample=True, top_p=0.9, temperature=0.7, repetition_penalty=2.5)
print(tokenizer.decode(output[0], skip_special_tokens=True))