Flux Prompt Enhance

🚀 Flux-Prompt-Enhance

Flux-Prompt-Enhance 是一個文本生成模型，它基於 google-t5/t5-base 模型，可用於將簡短的提示詞轉換為更詳細、豐富的描述，適用於文本到文本的生成任務。

🚀 快速開始

以下是使用 Flux-Prompt-Enhance 模型進行提示詞增強的示例代碼：

from transformers import pipeline, AutoTokenizer, AutoModelForSeq2SeqLM

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

# Model checkpoint
model_checkpoint = "gokaygokay/Flux-Prompt-Enhance"

# Tokenizer
tokenizer = AutoTokenizer.from_pretrained(model_checkpoint)

# Model
model = AutoModelForSeq2SeqLM.from_pretrained(model_checkpoint)

enhancer = pipeline('text2text-generation',
                    model=model,
                    tokenizer=tokenizer,
                    repetition_penalty= 1.2,
                    device=device)

max_target_length = 256
prefix = "enhance prompt: "

short_prompt = "beautiful house with text 'hello'"
answer = enhancer(prefix + short_prompt, max_length=max_target_length)
final_answer = answer[0]['generated_text']
print(final_answer)

# a two-story house with white trim, large windows on the second floor,
# three chimneys on the roof, green trees and shrubs in front of the house,
# stone pathway leading to the front door, text on the house reads "hello" in all caps,
# blue sky above, shadows cast by the trees, sunlight creating contrast on the house's facade,
# some plants visible near the bottom right corner, overall warm and serene atmosphere. 

💻 使用示例

基礎用法

from transformers import pipeline, AutoTokenizer, AutoModelForSeq2SeqLM

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

# Model checkpoint
model_checkpoint = "gokaygokay/Flux-Prompt-Enhance"

# Tokenizer
tokenizer = AutoTokenizer.from_pretrained(model_checkpoint)

# Model
model = AutoModelForSeq2SeqLM.from_pretrained(model_checkpoint)

enhancer = pipeline('text2text-generation',
                    model=model,
                    tokenizer=tokenizer,
                    repetition_penalty= 1.2,
                    device=device)

max_target_length = 256
prefix = "enhance prompt: "

short_prompt = "beautiful house with text 'hello'"
answer = enhancer(prefix + short_prompt, max_length=max_target_length)
final_answer = answer[0]['generated_text']
print(final_answer)

高級用法

# 可以根據具體需求調整參數，如 max_target_length、repetition_penalty 等，以獲得不同的生成效果
from transformers import pipeline, AutoTokenizer, AutoModelForSeq2SeqLM

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

# Model checkpoint
model_checkpoint = "gokaygokay/Flux-Prompt-Enhance"

# Tokenizer
tokenizer = AutoTokenizer.from_pretrained(model_checkpoint)

# Model
model = AutoModelForSeq2SeqLM.from_pretrained(model_checkpoint)

enhancer = pipeline('text2text-generation',
                    model=model,
                    tokenizer=tokenizer,
                    repetition_penalty= 1.5,  # 調整重複懲罰係數
                    device=device)

max_target_length = 300  # 增加最大目標長度
prefix = "enhance prompt: "

short_prompt = "beautiful house with text 'hello'"
answer = enhancer(prefix + short_prompt, max_length=max_target_length)
final_answer = answer[0]['generated_text']
print(final_answer)

📄 許可證

本項目採用 apache-2.0 許可證。

📚 詳細文檔

模型信息