import gradio as gr import io import random import time import json import base64 import requests import os from mistralai import Mistral from PIL import Image from io import BytesIO from deep_translator import GoogleTranslator from datetime import datetime from theme import theme from fastapi import FastAPI app = FastAPI() API_URL = "https://api-inference.huggingface.co/models/stabilityai/stable-diffusion-3.5-large-turbo" API_TOKEN = os.getenv("HF_READ_TOKEN") headers = {"Authorization": f"Bearer {API_TOKEN}"} timeout = 100 api_key = os.getenv("MISTRAL_API_KEY") Mistralclient = Mistral(api_key=api_key) def clear(): return None def change_tab(): return gr.Tabs.update(selected=1) # Function to query the API and return the generated image def query(prompt, is_negative=False, steps=35, cfg_scale=7, sampler="DPM++ 2M Karras", seed=-1, strength=0.7, width=896, height=1152): if prompt == "" or prompt is None: return None key = random.randint(0, 999) API_TOKEN = random.choice([os.getenv("HF_READ_TOKEN")]) headers = {"Authorization": f"Bearer {API_TOKEN}"} # Translate the prompt from Russian to English if necessary prompt = GoogleTranslator(source='ru', target='en').translate(prompt) print(f'\033[1mGeneration {key} translation:\033[0m {prompt}') # Add some extra flair to the prompt prompt = f"{prompt} | ultra detail, ultra elaboration, ultra quality, perfect." print(f'\033[1mGeneration {key}:\033[0m {prompt}') # Prepare the payload for the API call, including width and height payload = { "inputs": prompt, "is_negative": is_negative, "steps": steps, "cfg_scale": cfg_scale, "seed": seed if seed != -1 else random.randint(1, 1000000000), "strength": strength, "parameters": { "width": width, # Pass the width to the API "height": height # Pass the height to the API } } # Send the request to the API and handle the response response = requests.post(API_URL, headers=headers, json=payload, timeout=timeout) if response.status_code != 200: print(f"Error: Failed to get image. Response status: {response.status_code}") print(f"Response content: {response.text}") if response.status_code == 503: raise gr.Error(f"{response.status_code} : The model is being loaded") raise gr.Error(f"{response.status_code}") try: # Convert the response content into an image image_bytes = response.content image = Image.open(io.BytesIO(image_bytes)) print(f'\033[1mGeneration {key} completed!\033[0m ({prompt})') return image except Exception as e: print(f"Error when trying to open the image: {e}") return None examples = [ "a beautiful woman with blonde hair and blue eyes", "a beautiful woman with brown hair and grey eyes", "a beautiful woman with black hair and brown eyes", ] def encode_image(image_path): """Encode the image to base64.""" try: # Open the image file image = Image.open(image_path).convert("RGB") # Resize the image to a height of 512 while maintaining the aspect ratio base_height = 512 h_percent = (base_height / float(image.size[1])) w_size = int((float(image.size[0]) * float(h_percent))) image = image.resize((w_size, base_height), Image.LANCZOS) # Convert the image to a byte stream buffered = BytesIO() image.save(buffered, format="JPEG") img_str = base64.b64encode(buffered.getvalue()).decode("utf-8") return img_str except FileNotFoundError: print(f"Error: The file {image_path} was not found.") return None except Exception as e: # Add generic exception handling print(f"Error: {e}") return None def feifeichat(image): try: model = "pixtral-large-2411" # Define the messages for the chat base64_image = encode_image(image) messages = [{ "role": "user", "content": [ { "type": "text", "text": "Please provide a detailed description of this photo" }, { "type": "image_url", "image_url": f"data:image/jpeg;base64,{base64_image}" }, ], "stream": False, }] partial_message = "" for chunk in Mistralclient.chat.stream(model=model, messages=messages): if chunk.data.choices[0].delta.content is not None: partial_message = partial_message + chunk.data.choices[ 0].delta.content yield partial_message except Exception as e: # 添加通用异常处理 print(f"Error: {e}") return "Please upload a photo" # CSS to style the app css = """ footer{display:none !important} #app-container { max-width: 930px; margin-left: auto; margin-right: auto; } """ # Build the Gradio UI with Blocks with gr.Blocks(theme=theme, css=css) as app: # Add a title to the app gr.HTML("

🎨 Stable Diffusion 3.5 large turbo 🇬🇧

") with gr.Tabs() as tabs: with gr.TabItem(label="Image To Prompt", visible=True, id=1): with gr.Row(): with gr.Column(): input_img = gr.Image(label="Input Picture",height=320,type="filepath") submit_btn = gr.Button(value="Submit", variant='primary') with gr.Column(): output_text = gr.Textbox(label="Flux Prompt", show_copy_button = True) clr_button =gr.Button("Clear",variant="primary", elem_id="clear_button") clr_button.click(lambda: gr.Textbox(value=""), None, output_text) submit_btn.click(feifeichat, [input_img], [output_text]) with gr.TabItem("Text to Image", visible=True, id=0): # Container for all the UI elements with gr.Column(elem_id="app-container"): # Add a text input for the main prompt with gr.Row(): with gr.Column(elem_id="prompt-container"): with gr.Group(): with gr.Row(): text_prompt = gr.Textbox(label="Image Prompt ✍️", placeholder="Enter a prompt here", lines=2, show_copy_button = True, elem_id="prompt-text-input") # Accordion for advanced settings with gr.Row(): with gr.Accordion("Advanced Settings", open=False): negative_prompt = gr.Textbox(label="Negative Prompt", lines=4, placeholder="What should not be in the image", value="(hands:-1.25), physical-defects:2, unhealthy-deformed-joints:2, unhealthy-hands:2, out of frame, bad face, (bad-image-v2-39000:1.3), (((out of frame))), deformed body features, poor facial details, (poorly drawn face:1.3), jpeg artifacts, (missing arms:1.1), (missing legs:1.1), (extra arms:1.2), (extra legs:1.2)") with gr.Row(): width = gr.Slider(label="ImageWidth", value=896, minimum=64, maximum=1216, step=32) height = gr.Slider(label="Image Height", value=1152, minimum=64, maximum=1216, step=32) steps = gr.Slider(label="Sampling steps", value=50, minimum=1, maximum=100, step=1) cfg = gr.Slider(label="CFG Scale", value=3.5, minimum=1, maximum=20, step=1) strength = gr.Slider(label="PromptStrength", value=100, minimum=0, maximum=100, step=1) seed = gr.Slider(label="Seed", value=-1, minimum=-1, maximum=1000000000, step=1) # Setting the seed to -1 will make it random method = gr.Radio(label="Sampling method", value="DPM++ 2M Karras", choices=["DPM++ 2M Karras", "DPM++ 2S a Karras", "DPM2 a Karras", "DPM2 Karras", "DPM++ SDE Karras", "DEIS", "LMS", "DPM Adaptive", "DPM++ 2M", "DPM2 Ancestral", "DPM++ S", "DPM++ SDE", "DDPM", "DPM Fast", "dpmpp_2s_ancestral", "Euler", "Euler CFG PP", "Euler a", "Euler Ancestral", "Euler+beta", "Heun", "Heun PP2", "DDIM", "LMS Karras", "PLMS", "UniPC", "UniPC BH2"]) # Add a button to trigger the image generation with gr.Row(): text_button = gr.Button("Generate Image 🎨", variant='primary', elem_id="gen-button") clear_prompt =gr.Button("Clear Prompt 🗑️",variant="primary", elem_id="clear_button") clear_prompt.click(lambda: (None), None, [text_prompt], queue=False, show_api=False) with gr.Group(): # Image output area to display the generated image with gr.Row(): image_output = gr.Image(type="pil", label="Image Output", format="png", show_share_button=False, elem_id="gallery") gr.Examples( examples = examples, inputs = [text_prompt], ) with gr.Row(): clear_results = gr.Button(value="Clear Image 🗑️", variant="primary", elem_id="clear_button") clear_results.click(lambda: (None), None, [image_output], queue=False, show_api=False) # Bind the button to the query function with the added width and height inputs text_button.click(query, inputs=[text_prompt, negative_prompt, steps, cfg, method, seed, strength, width, height], outputs=image_output) app.queue(default_concurrency_limit=200, max_size=200) # <-- Sets up a queue with default parameters if __name__ == "__main__": app.launch(show_api=False, share=False)