import gradio as gr import requests import time import random import json import base64 import os from transformers import pipeline, set_seed from io import BytesIO import io import html import PIL from PIL import Image import re def query(payload, model): HF_TOKEN = os.getenv("HF_TOKEN") headers = {"Authorization": f"Bearer {HF_TOKEN}"} url = "https://api-inference.huggingface.co/models/" API_URL = f"{url}{model}" response = requests.post(API_URL, headers=headers, json=payload) return response.content def hf_inference(prompt, negative, model, steps, sampler, guidance, width, height, seed, progress=gr.Progress()): try: progress(0, desc="Starting") images=[] time.sleep(2.5) progress(0.05) progress(0.25, desc="Generating") image_bytes = query(payload={ "inputs": f"{prompt}", "parameters": { "negative_prompt": f"{negative}", "num_inference_steps": steps, "guidance_scale": guidance, "width": width, "height": height, "seed": seed, }, }, model=model) progress(0.75, desc="Opening image") time.sleep(1) image = Image.open(io.BytesIO(image_bytes)) images.append(image) progress(0.99, desc="Sending image") time.sleep(0.5) return images except PIL.UnidentifiedImageError: gr.Warning("This model is not loaded now. Try others models.") class Prodia: def __init__(self, api_key, base=None): self.base = base or "https://api.prodia.com/v1" self.headers = { "X-Prodia-Key": api_key } def generate(self, params): response = self._post(f"{self.base}/sd/generate", params) return response.json() def transform(self, params): response = self._post(f"{self.base}/sd/transform", params) return response.json() def controlnet(self, params): response = self._post(f"{self.base}/sd/controlnet", params) return response.json() def get_job(self, job_id): response = self._get(f"{self.base}/job/{job_id}") return response.json() def wait(self, job): job_result = job while job_result['status'] not in ['succeeded', 'failed']: time.sleep(0.25) job_result = self.get_job(job['job']) return job_result def list_models(self): response = self._get(f"{self.base}/sd/models") return response.json() def list_samplers(self): response = self._get(f"{self.base}/sd/samplers") return response.json() def _post(self, url, params): headers = { **self.headers, "Content-Type": "application/json" } response = requests.post(url, headers=headers, data=json.dumps(params)) if response.status_code != 200: raise Exception(f"Bad Prodia Response: {response.status_code}") return response def _get(self, url): response = requests.get(url, headers=self.headers) if response.status_code != 200: raise Exception(f"Bad Prodia Response: {response.status_code}") return response def image_to_base64(image): # Convert the image to bytes buffered = BytesIO() image.save(buffered, format="PNG") # You can change format to PNG if needed # Encode the bytes to base64 img_str = base64.b64encode(buffered.getvalue()) return img_str.decode('utf-8') # Convert bytes to string def remove_id_and_ext(text): text = re.sub(r'\[.*\]$', '', text) extension = text[-12:].strip() if extension == "safetensors": text = text[:-13] elif extension == "ckpt": text = text[:-4] return text def get_data(text): results = {} patterns = { 'prompt': r'(.*)', 'negative_prompt': r'Negative prompt: (.*)', 'steps': r'Steps: (\d+),', 'seed': r'Seed: (\d+),', 'sampler': r'Sampler:\s*([^\s,]+(?:\s+[^\s,]+)*)', 'model': r'Model:\s*([^\s,]+)', 'cfg_scale': r'CFG scale:\s*([\d\.]+)', 'size': r'Size:\s*([0-9]+x[0-9]+)' } for key in ['prompt', 'negative_prompt', 'steps', 'seed', 'sampler', 'model', 'cfg_scale', 'size']: match = re.search(patterns[key], text) if match: results[key] = match.group(1) else: results[key] = None if results['size'] is not None: w, h = results['size'].split("x") results['w'] = w results['h'] = h else: results['w'] = None results['h'] = None return results def send_to_img2img_def(images): return images def send_to_txt2img(image): result = {tabs: gr.update(selected="t2i")} try: text = image.info['parameters'] data = get_data(text) result[prompt] = gr.update(value=data['prompt']) result[negative_prompt] = gr.update(value=data['negative_prompt']) if data[ 'negative_prompt'] is not None else gr.update() result[steps] = gr.update(value=int(data['steps'])) if data['steps'] is not None else gr.update() result[seed] = gr.update(value=int(data['seed'])) if data['seed'] is not None else gr.update() result[cfg_scale] = gr.update(value=float(data['cfg_scale'])) if data['cfg_scale'] is not None else gr.update() result[width] = gr.update(value=int(data['w'])) if data['w'] is not None else gr.update() result[height] = gr.update(value=int(data['h'])) if data['h'] is not None else gr.update() result[sampler] = gr.update(value=data['sampler']) if data['sampler'] is not None else gr.update() if model in model_names: result[model] = gr.update(value=model_names[model]) else: result[model] = gr.update() return result except Exception as e: print(e) return result prodia_client = Prodia(api_key=os.getenv("PRODIA_API_KEY")) model_list = prodia_client.list_models() model_names = {} for model_name in model_list: name_without_ext = remove_id_and_ext(model_name) model_names[name_without_ext] = model_name def txt2img(prompt, negative_prompt, model, steps, sampler, cfg_scale, width, height, seed, progress=gr.Progress()): progress(0, desc="Starting") time.sleep(2.5) progress(0.25, desc="Generating") result = prodia_client.generate({ "prompt": prompt, "negative_prompt": negative_prompt, "model": model, "steps": steps, "sampler": sampler, "cfg_scale": cfg_scale, "width": width, "height": height, "seed": seed }) progress(0.75, desc="Opening image") job = prodia_client.wait(result) progress(0.99, desc="Sending image") return [job["imageUrl"]], job["imageUrl"] def img2img(input_image, denoising, prompt, negative_prompt, model, steps, sampler, cfg_scale, width, height, seed, progress=gr.Progress()): progress(0, desc="Starting") time.sleep(1.5) progress(0.10, desc="Uploading input image") time.sleep(1.5) progress(0.25, desc="Generating") result = prodia_client.transform({ "imageData": image_to_base64(input_image), "denoising_strength": denoising, "prompt": prompt, "negative_prompt": negative_prompt, "model": model, "steps": steps, "sampler": sampler, "cfg_scale": cfg_scale, "width": width, "height": height, "seed": seed }) progress(0.75, desc="Opening image") job = prodia_client.wait(result) progress(0.99, desc="Sending image") time.sleep(0.5) return [job["imageUrl"]], job["imageUrl"] css = """ #generate { height: 100%; } """ with gr.Blocks(css=css) as demo: with gr.Row(): with gr.Column(scale=6): model = gr.Dropdown(interactive=True, value="absolutereality_v181.safetensors [3d9d4d2b]", show_label=True, label="Stable Diffusion Checkpoint", choices=prodia_client.list_models()) with gr.Tabs() as tabs: with gr.Tab("txt2img", id='t2i'): with gr.Row(): with gr.Column(scale=6, min_width=600): prompt = gr.Textbox(placeholder="Prompt", show_label=False, lines=3) negative_prompt = gr.Textbox(placeholder="Negative Prompt", show_label=False, lines=3, value="(deformed, distorted, disfigured:1.3), poorly drawn, bad anatomy, wrong anatomy, extra limb, missing limb, floating limbs, (mutated hands and fingers:1.4), disconnected limbs, mutation, mutated, ugly, disgusting, blurry, amputation") with gr.Column(): text_button = gr.Button("Generate", variant='primary', elem_id="generate") with gr.Row(): with gr.Column(scale=3): with gr.Tab("Generation"): with gr.Row(): with gr.Column(scale=1): sampler = gr.Dropdown(value="DPM++ 2M Karras", show_label=True, label="Sampling Method", choices=prodia_client.list_samplers()) with gr.Column(scale=1): steps = gr.Slider(label="Sampling Steps", minimum=1, maximum=25, value=20, step=1) with gr.Row(): with gr.Column(scale=1): width = gr.Slider(label="Width", maximum=1024, value=512, step=8) height = gr.Slider(label="Height", maximum=1024, value=512, step=8) with gr.Column(scale=1): batch_size = gr.Slider(label="Batch Size", maximum=1, value=1) batch_count = gr.Slider(label="Batch Count", maximum=1, value=1) cfg_scale = gr.Slider(label="CFG Scale", minimum=1, maximum=20, value=8, step=1) seed = gr.Number(label="Seed", value=-1) with gr.Column(scale=2): image_output = gr.Gallery(show_label=False, rows=2) send_to_img2img = gr.Button(value="Send OUTPUT IMAGE to img2img") send_to_png = gr.Button(value="Send OUTPUT IMAGE to PNG Info") past_url = gr.Textbox(visible=False, interactive=False) text_button.click(txt2img, inputs=[prompt, negative_prompt, model, steps, sampler, cfg_scale, width, height, seed], outputs=[image_output, past_url], concurrency_limit=64) with gr.Tab("img2img", id='i2i'): with gr.Row(): with gr.Column(scale=6, min_width=600): i2i_prompt = gr.Textbox(placeholder="Prompt", show_label=False, lines=3) i2i_negative_prompt = gr.Textbox(placeholder="Negative Prompt", show_label=False, lines=3, value="(deformed, distorted, disfigured:1.3), poorly drawn, bad anatomy, wrong anatomy, extra limb, missing limb, floating limbs, (mutated hands and fingers:1.4), disconnected limbs, mutation, mutated, ugly, disgusting, blurry, amputation") with gr.Column(): i2i_text_button = gr.Button("Generate", variant='primary', elem_id="generate") with gr.Row(): with gr.Column(scale=3): with gr.Tab("Generation"): i2i_image_input = gr.Image(type="pil", interactive=True) with gr.Row(): with gr.Column(scale=1): i2i_sampler = gr.Dropdown(value="Euler a", show_label=True, label="Sampling Method", choices=prodia_client.list_samplers()) with gr.Column(scale=1): i2i_steps = gr.Slider(label="Sampling Steps", minimum=1, maximum=25, value=20, step=1) with gr.Row(): with gr.Column(scale=1): i2i_width = gr.Slider(label="Width", maximum=1024, value=512, step=8) i2i_height = gr.Slider(label="Height", maximum=1024, value=512, step=8) with gr.Column(scale=1): i2i_batch_size = gr.Slider(label="Batch Size", maximum=1, value=1) i2i_batch_count = gr.Slider(label="Batch Count", maximum=1, value=1) i2i_cfg_scale = gr.Slider(label="CFG Scale", minimum=1, maximum=20, value=7, step=1) i2i_denoising = gr.Slider(label="Denoising Strength", minimum=0, maximum=1, value=0.7, step=0.1) i2i_seed = gr.Number(label="Seed", value=-1) with gr.Column(scale=2): i2i_image_output = gr.Gallery(show_label=False, rows=2) send_to_png_i2i = gr.Button(value="Send INPUT IMAGE to PNG Info") i2i_past_url = gr.Textbox(visible=False, interactive=False) i2i_text_button.click(img2img, inputs=[i2i_image_input, i2i_denoising, i2i_prompt, i2i_negative_prompt, model, i2i_steps, i2i_sampler, i2i_cfg_scale, i2i_width, i2i_height, i2i_seed], outputs=[i2i_image_output, i2i_past_url], concurrency_limit=64) send_to_img2img.click(send_to_img2img_def, inputs=past_url, outputs=i2i_image_input) with gr.Tab("PNG Info"): def plaintext_to_html(text, classname=None): content = "
\n".join(html.escape(x) for x in text.split('\n')) return f"

{content}

" if classname else f"

{content}

" def get_exif_data(path): image = Image.open(path) items = image.info info = '' for key, text in items.items(): info += f"""

{plaintext_to_html(str(key))}

{plaintext_to_html(str(text))}

""".strip() + "\n" if len(info) == 0: message = "Nothing found in the image." info = f"

{message}

" return info with gr.Row(): with gr.Column(): image_input = gr.Image(type="filepath", interactive=True) png_button = gr.Button("Get Info") with gr.Row(): with gr.Column(): exif_output = gr.HTML(label="EXIF Data") send_to_txt2img_btn = gr.Button("Send PARAMETRS to txt2img") send_to_img2img_png = gr.Button("Send IMAGE to img2img") image_input.upload(get_exif_data, inputs=[image_input], outputs=exif_output) png_button.click(get_exif_data, inputs=[image_input], outputs=exif_output) send_to_txt2img_btn.click(send_to_txt2img, inputs=[image_input], outputs=[tabs, prompt, negative_prompt, steps, seed, model, sampler, width, height, cfg_scale], concurrency_limit=64) send_to_png.click(send_to_img2img_def, inputs=past_url, outputs=image_input) send_to_img2img_png.click(send_to_img2img_def, inputs=past_url, outputs=i2i_image_input) send_to_png_i2i.click(send_to_img2img_def, inputs=i2i_past_url, outputs=image_input) with gr.Tab("HuggingFace Inference"): with gr.Row(): gr.Markdown("Add your model from HF.co, enter model ID.") hf_model = gr.Dropdown(label="HuggingFace checkpoint", choices=["runwayml/stable-diffusion-v1-5", "stabilityai/stable-diffusion-2-1", "dataautogpt3/OpenDalleV1.1", "CompVis/stable-diffusion-v1-4", "playgroundai/playground-v2-1024px-aesthetic", "prompthero/openjourney", "openskyml/dreamdrop-v1", "SG161222/Realistic_Vision_V1.4", "digiplay/AbsoluteReality_v1.8.1", "openskyml/dalle-3-xl", "Lykon/dreamshaper-7", "Pclanglais/Mickey-1928"], value="runwayml/stable-diffusion-v1-5", allow_custom_value=True, interactive=True) with gr.Row(): with gr.Column(scale=6, min_width=600): hf_prompt = gr.Textbox(placeholder="Prompt", show_label=False, lines=3) hf_negative_prompt = gr.Textbox(placeholder="Negative Prompt", show_label=False, lines=3, value="(deformed, distorted, disfigured:1.3), poorly drawn, bad anatomy, wrong anatomy, extra limb, missing limb, floating limbs, (mutated hands and fingers:1.4), disconnected limbs, mutation, mutated, ugly, disgusting, blurry, amputation") with gr.Column(): hf_text_button = gr.Button("Generate with HF", variant='primary', elem_id="generate") with gr.Row(): with gr.Column(scale=3): with gr.Tab("Generation"): with gr.Row(): with gr.Column(scale=1): hf_steps = gr.Slider(label="Sampling Steps", minimum=1, maximum=25, value=20, step=1) with gr.Row(): with gr.Column(scale=1): hf_width = gr.Slider(label="Width", maximum=1024, value=512, step=8) hf_height = gr.Slider(label="Height", maximum=1024, value=512, step=8) with gr.Column(scale=1): hf_batch_size = gr.Slider(label="Batch Size", maximum=1, value=1) hf_batch_count = gr.Slider(label="Batch Count", maximum=1, value=1) hf_cfg_scale = gr.Slider(label="CFG Scale", minimum=1, maximum=20, value=8, step=1) hf_seed = gr.Number(label="Seed", value=-1) with gr.Column(scale=2): hf_image_output = gr.Gallery(show_label=False, preview=True, columns=4, allow_preview=True) #hf_send_to_img2img = gr.Button(value="Send to img2img") hf_text_button.click(hf_inference, inputs=[hf_prompt, hf_negative_prompt, hf_model, hf_steps, sampler, hf_cfg_scale, hf_width, hf_height, hf_seed], outputs=hf_image_output, concurrency_limit=64) with gr.Tab("Prompt Generator"): gpt2_pipe = pipeline('text-generation', model='Gustavosta/MagicPrompt-Stable-Diffusion', tokenizer='gpt2') with open("ideas.txt", "r") as f: line = f.readlines() def generate(starting_text): seed = random.randint(100, 1000000) set_seed(seed) if starting_text == "": starting_text: str = line[random.randrange(0, len(line))].replace("\n", "").lower().capitalize() starting_text: str = re.sub(r"[,:\-–.!;?_]", '', starting_text) response = gpt2_pipe(starting_text, max_length=(len(starting_text) + random.randint(60, 90)), num_return_sequences=4) response_list = [] for x in response: resp = x['generated_text'].strip() if resp != starting_text and len(resp) > (len(starting_text) + 4) and resp.endswith((":", "-", "—")) is False: response_list.append(resp+'\n') response_end = "\n".join(response_list) response_end = re.sub('[^ ]+\.[^ ]+','', response_end) response_end = response_end.replace("<", "").replace(">", "") if response_end != "": return response_end with gr.Tab("BLIP"): with gr.Tab("Base"): gr.load("models/Salesforce/blip-image-captioning-base", title="BLIP-base") with gr.Tab("Large"): gr.load("models/Salesforce/blip-image-captioning-large", title="BLIP-large") with gr.Tab("Classification"): gr.load("models/google/vit-base-patch16-224", title="ViT Classification") #with gr.Tab("Segmentation"): # gr.load("models/mattmdjaga/segformer_b2_clothes", title="SegFormer Segmentation") with gr.Tab("Visual Question Answering"): gr.load("models/dandelin/vilt-b32-finetuned-vqa", title="ViLT VQA") demo.queue(max_size=80, api_open=False).launch(max_threads=256, show_api=False)