import gradio as gr
from huggingface_hub import InferenceClient
from gradio_client import Client
from PIL import Image
import requests
from io import BytesIO
import streamlit as st
# Initialize the HuggingFace Inference Client with the specified model
client_mistral = InferenceClient("mistralai/Mistral-7B-Instruct-v0.2")
# Initialize the Playground AI client
client_playground = Client("https://playgroundai-playground-v2-5.hf.space/--replicas/c9ozb/")
def format_prompt(logo_request):
system_prompt = """
You are an advanced language model designed to create detailed and creative image prompts for logo generation. Based on the user's input, generate an elaborate and descriptive image prompt that can be used to create a high-quality logo. Ensure that the prompt is clear, imaginative, and provides specific details that will guide the logo creation process effectively.
"""
prompt = f"[SYS] {system_prompt} [/SYS][INST] {logo_request} [/INST]"
return prompt
def generate_improved_prompt(logo_request, temperature=0.9, max_new_tokens=512, top_p=0.95, repetition_penalty=1.0):
temperature = float(temperature)
if temperature < 1e-2:
temperature = 1e-2
top_p = float(top_p)
generate_kwargs = {
"temperature": temperature,
"max_new_tokens": max_new_tokens,
"top_p": top_p,
"repetition_penalty": repetition_penalty,
"do_sample": True,
"seed": 42,
}
formatted_prompt = format_prompt(logo_request)
stream = client_mistral.text_generation(formatted_prompt, **generate_kwargs, stream=True, details=True, return_full_text=False)
output = ""
for response in stream:
output += response.token.text
yield output
def generate_image(prompt, negative_prompt="", use_negative_prompt=False, seed=0, width=1024, height=1024, guidance_scale=7.5, randomize_seed=True):
result = client_playground.predict(
prompt,
negative_prompt,
use_negative_prompt,
seed,
width,
height,
guidance_scale,
randomize_seed,
api_name="/run"
)
# Extract the image URL from the result
image_path = result[0][0]["image"]
image_url = "https://playgroundai-playground-v2-5.hf.space/--replicas/c9ozb/file=" + image_path
# Fetch and display the result image
response = requests.get(image_url)
if response.headers['Content-Type'].startswith('image'):
img = Image.open(BytesIO(response.content))
return img
else:
return None
css = """
#mkd {
height: 500px;
overflow: auto;
border: 1px solid #ccc;
}
"""
with gr.Blocks(css=css) as gpt:
with gr.Row():
with gr.Column(scale=2):
gr.HTML("