import gradio as gr import requests import io import random import os from PIL import Image from huggingface_hub import InferenceClient from deep_translator import GoogleTranslator from gradio_client import Client # os.makedirs('assets', exist_ok=True) if not os.path.exists('icon.png'): os.system("wget -O icon.png https://i.pinimg.com/564x/64/49/88/644988c59447eb00286834c2e70fdd6b.jpg") API_URL_DEV = "https://api-inference.huggingface.co/models/black-forest-labs/FLUX.1-dev" API_URL = "https://api-inference.huggingface.co/models/black-forest-labs/FLUX.1-schnell" timeout = 100 def enhance_prompt(prompt, model="mistralai/Mistral-7B-Instruct-v0.1", style="photo-realistic"): client = Client("K00B404/Mistral-Nemo-custom") system_prompt=f""" You are a image generation prompt enhancer specialized in the {style} style. You must respond only with the enhanced version of the users input prompt Remember, image generation models can be stimulated by refering to camera 'effect' in the prompt like :4k ,award winning, super details, 35mm lens, hd """ user_message=f"###input image generation prompt### {prompt}" result = client.predict( system_prompt=system_prompt, user_message=user_message, max_tokens=256, model_id=model,# "mistralai/Mistral-Nemo-Instruct-2407", api_name="/predict" ) return result # The output value that appears in the "Response" Textbox component. """result = client.predict( system_prompt=system_prompt,#"You are a image generation prompt enhancer and must respond only with the enhanced version of the users input prompt", user_message=user_message, max_tokens=500, api_name="/predict" ) return result """ def enhance_prompt_v2(prompt, model="mistralai/Mistral-Nemo-Instruct-2407", style="photo-realistic"): client = Client("K00B404/Mistral-Nemo-custom") system_prompt=f""" You are a image generation prompt enhancer specialized in the {style} style. You must respond only with the enhanced version of the users input prompt Remember, image generation models can be stimulated by refering to camera 'effect' in the prompt like :4k ,award winning, super details, 35mm lens, hd """ user_message=f"###input image generation prompt### {prompt}" result = client.predict( system_prompt=system_prompt, user_message=user_message, max_tokens=256, model_id=model, api_name="/predict" ) return result def mistral_nemo_call(prompt, API_TOKEN, model="mistralai/Mistral-Nemo-Instruct-2407", style="photo-realistic"): client = InferenceClient(api_key=API_TOKEN) system_prompt=f""" You are a image generation prompt enhancer specialized in the {style} style. You must respond only with the enhanced version of the users input prompt Remember, image generation models can be stimulated by refering to camera 'effect' in the prompt like :4k ,award winning, super details, 35mm lens, hd """ response = "" for message in client.chat_completion( model=model, messages=[{"role": "system", "content": system_prompt}, {"role": "user", "content": prompt} ], max_tokens=500, stream=True, ): response += message.choices[0].delta.content return response def query(prompt, is_negative=False, steps=30, cfg_scale=7, sampler="DPM++ 2M Karras", seed=-1, strength=0.7, huggingface_api_key=None, use_dev=False,enhance_prompt_style="generic", enhance_prompt_option=False, nemo_enhance_prompt_style="generic", use_mistral_nemo=False): # Determine which API URL to use api_url = API_URL_DEV if use_dev else API_URL # Check if the request is an API call by checking for the presence of the huggingface_api_key is_api_call = huggingface_api_key is not None if is_api_call: # Use the environment variable for the API key in GUI mode API_TOKEN = os.getenv("HF_READ_TOKEN") else: # Validate the API key if it's an API call if huggingface_api_key == "": raise gr.Error("API key is required for API calls.") API_TOKEN = huggingface_api_key headers = {"Authorization": f"Bearer {API_TOKEN}"} if prompt == "" or prompt is None: return None, None, None key = random.randint(0, 999) prompt = GoogleTranslator(source='ru', target='en').translate(prompt) print(f'\033[1mGeneration {key} translation:\033[0m {prompt}') original_prompt = prompt if enhance_prompt_option: prompt = enhance_prompt_v2(prompt, style=enhance_prompt_style) print(f'\033[1mGeneration {key} enhanced prompt:\033[0m {prompt}') if use_mistral_nemo: prompt = mistral_nemo_call(prompt, API_TOKEN=API_TOKEN, style=nemo_enhance_prompt_style) print(f'\033[1mGeneration {key} Mistral-Nemo prompt:\033[0m {prompt}') final_prompt = f"{prompt} | ultra detail, ultra elaboration, ultra quality, perfect." print(f'\033[1mGeneration {key}:\033[0m {final_prompt}') # If seed is -1, generate a random seed and use it if seed == -1: seed = random.randint(1, 1000000000) payload = { "inputs": final_prompt, "is_negative": is_negative, "steps": steps, "cfg_scale": cfg_scale, "seed": seed, "strength": strength } 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: image_bytes = response.content image = Image.open(io.BytesIO(image_bytes)) print(f'\033[1mGeneration {key} completed!\033[0m ({final_prompt})') # Save the image to a file and return the file path and seed output_path = f"./output_{key}.png" image.save(output_path) return output_path, seed, prompt if enhance_prompt_option else original_prompt except Exception as e: print(f"Error when trying to open the image: {e}") return None, None, None css = """ body { background-image: url('icon.png'); background-size: cover; background-repeat: no-repeat; background-attachment: fixed; } #app-container { background-color: rgba(0, 0, 0, 0.001); /* semi-transparent white */ max-width: 600px; margin-left: auto; margin-right: auto; padding: 20px; border-radius: 10px; box-shadow: 0 0 10px rgba(0,0,0,0.001); } #title-container { display: flex; align-items: center; justify-content: center; } #title-icon { width: 32px; height: auto; margin-right: 10px; } #title-text { font-size: 24px; font-weight: bold; } """ css1 = """ #app-container { max-width: 600px; margin-left: auto; margin-right: auto; } #title-container { display: flex; align-items: center; justify-content: center; } #app-container { max-width: 600px; margin-left: auto; margin-right: auto; background-color: rgba(255, 255, 255, 0.001); /* semi-transparent white */ padding: 20px; border-radius: 10px; } #title-icon { width: 32px; /* Adjust the width of the icon as needed */ height: auto; margin-right: 10px; /* Space between icon and title */ } #title-text { font-size: 24px; /* Adjust font size as needed */ font-weight: bold; } """ with gr.Blocks(theme='Nymbo/Nymbo_Theme', css=css) as app: gr.HTML("""

FLUX Capacitor

""") with gr.Column(elem_id="app-container"): with gr.Row(): with gr.Column(elem_id="prompt-container"): with gr.Row(): text_prompt = gr.Textbox(label="Prompt", placeholder="Enter a prompt here", lines=2, elem_id="prompt-text-input") with gr.Row(): with gr.Accordion("Advanced Settings", open=False): negative_prompt = gr.Textbox(label="Negative Prompt", placeholder="What should not be in the image", value="(deformed, distorted, disfigured), poorly drawn, bad anatomy, wrong anatomy, extra limb, missing limb, floating limbs, (mutated hands and fingers), disconnected limbs, mutation, mutated, ugly, disgusting, blurry, amputation, misspellings, typos", lines=3, elem_id="negative-prompt-text-input") steps = gr.Slider(label="Sampling steps", value=35, minimum=1, maximum=100, step=1) cfg = gr.Slider(label="CFG Scale", value=7, minimum=1, maximum=20, step=1) method = gr.Radio(label="Sampling method", value="DPM++ 2M Karras", choices=["DPM++ 2M Karras", "DPM++ SDE Karras", "Euler", "Euler a", "Heun", "DDIM"]) strength = gr.Slider(label="Strength", value=0.7, minimum=0, maximum=1, step=0.001) seed = gr.Slider(label="Seed", value=-1, minimum=-1, maximum=1000000000, step=1) huggingface_api_key = gr.Textbox(label="Hugging Face API Key (required for API calls)", placeholder="Enter your Hugging Face API Key here", type="password", elem_id="api-key") use_dev = gr.Checkbox(label="Use Dev API", value=False, elem_id="use-dev-checkbox") enhance_prompt_style = gr.Textbox(label="Enhance Prompt Style", placeholder="Enter style for the prompt enhancer here", elem_id="enhance-prompt-style") enhance_prompt_option = gr.Checkbox(label="Enhance Prompt", value=False, elem_id="enhance-prompt-checkbox") use_mistral_nemo = gr.Checkbox(label="Use Mistral Nemo", value=False, elem_id="use-mistral-checkbox") nemo_prompt_style = gr.Textbox(label="Nemo Enhance Prompt Style", placeholder="Enter style for the prompt enhancer here", elem_id="nemo-enhance-prompt-style") with gr.Row(): text_button = gr.Button("Run", variant='primary', elem_id="gen-button") with gr.Row(): image_output = gr.Image(type="pil", label="Image Output", elem_id="gallery") with gr.Row(): seed_output = gr.Textbox(label="Seed Used", elem_id="seed-output") final_prompt_output = gr.Textbox(label="Final Prompt", elem_id="final-prompt-output") # Adjust the click function to include the API key, use_dev, and enhance_prompt_option as inputs text_button.click(query, inputs=[text_prompt, negative_prompt, steps, cfg, method, seed, strength, huggingface_api_key, use_dev, enhance_prompt_style,enhance_prompt_option, enhance_prompt_style, use_mistral_nemo], outputs=[image_output, seed_output, final_prompt_output]) app.launch(show_api=True, share=False)