File size: 3,173 Bytes
f667084 a04adbd 4d54b56 0372f7c fa3a39e a04adbd 264752e 50b4bab 264752e dcf269f 264752e 01b1364 0372f7c 3891dec a04adbd f667084 a04adbd f667084 50b4bab a04adbd 264752e 43c59b0 50b4bab dcf269f 50b4bab 3891dec fa3a39e f418994 dcf269f 264752e f418994 33407d1 9023169 a04adbd c83e28c 9023169 4d54b56 5ecb8ce f469466 252a605 01b1364 50b4bab 0c3147e 0276b97 0ef1f3f 264752e 43c59b0 264752e 0372f7c 264752e 9023169 f66faee 9023169 01a8184 0372f7c ae4c909 84bafee ae4c909 84bafee 257a388 4878b0f ca650cc 9023169 43c59b0 9023169 45289ed 9023169 0ef1f3f 9023169 a81e611 f667084 0ef1f3f |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 |
import os
from huggingface_hub import login
from transformers import BlipProcessor, BlipForConditionalGeneration
from PIL import Image
import gradio as gr
from diffusers import DiffusionPipeline
import torch
import spaces # Hugging Face Spaces module
import requests
from transformers import Qwen2VLForConditionalGeneration, AutoTokenizer, AutoProcessor
from qwen_vl_utils import process_vision_info
from diffusers import DiffusionPipeline
fabrics = ['cotton', 'silk', 'denim', 'linen', 'polyester', 'wool', 'velvet']
patterns = ['striped', 'floral', 'geometric', 'abstract', 'solid', 'polka dots']
textile_designs = ['woven texture', 'embroidery', 'printed fabric', 'hand-dyed', 'quilting']
# Get Hugging Face Token from environment variable
hf_token = os.getenv('HF_AUTH_TOKEN')
if not hf_token:
raise ValueError("Hugging Face token is not set in the environment variables.")
login(token=hf_token)
# Load the processor and model
processor = BlipProcessor.from_pretrained("Salesforce/blip-image-captioning-large")
model = BlipForConditionalGeneration.from_pretrained("Salesforce/blip-image-captioning-large")
processor1 = BlipProcessor.from_pretrained("noamrot/FuseCap")
model2 = BlipForConditionalGeneration.from_pretrained("noamrot/FuseCap")
pipe = DiffusionPipeline.from_pretrained("stabilityai/stable-diffusion-3.5-medium")
device = "cuda" if torch.cuda.is_available() else "cpu"
# pipe.to(device)
model2.to(device)
model.to(device)
pipe.to(device)
@spaces.GPU(duration=150)
def generate_caption_and_image(image):
img = image.convert("RGB")
# reader = easyocr.Reader(['en'])
# # result = reader.readtext(img)
import random
text = "a picture of "
inputs = processor(img, text, return_tensors="pt").to(device)
out = model2.generate(**inputs, num_beams = 3)
caption2 = processor.decode(out[0], skip_special_tokens=True)
# Generate caption
inputs = processor(image, return_tensors="pt", padding=True, truncation=True, max_length=250)
inputs = {key: val.to(device) for key, val in inputs.items()}
out = model.generate(**inputs)
caption1 = processor.decode(out[0], skip_special_tokens=True)
prompt = f"Generate a clothing item using the following details: 1. {caption1} 2. {caption2} 3. Fabric: {selected_fabric} 4. Pattern: {selected_pattern} 5. Design Style: {selected_textile_design}. "
prompt +="The image should have a clean, minimalistic grey or white background, with realistic lighting and fine details, ensuring a sophisticated and polished appearance"
# Generate image based on the caption
generated_image = pipe(prompt).images[0]
return prompt, generated_image
# Gradio UI
iface = gr.Interface(
fn=generate_caption_and_image,
inputs=[gr.Image(type="pil", label="Upload Image"),
gr.Radio(fabrics, label="Select Fabric"),
gr.Radio(patterns, label="Select Pattern"),
gr.Radio(textile_designs, label="Select Textile Design")]
outputs=[gr.Textbox(label="Generated Caption"), gr.Image(label="Generated Design")],
live=True
)
iface.launch(share=True)
|