LucyintheSky/lucy-feature-prediction

Feature Prediction

Model description

Predicts clothing features. Demo

"3/4 Sleeve", "Accessory", "Babydoll", "Closed Back", "Corset", "Crochet", "Cutouts", "Draped", "Floral", "Gloves", "Halter", "Lace", "Long", "Long Sleeve", "Midi", "No Slit", "Off The Shoulder", "One Shoulder", "Open Back", "Pockets", "Print", "Puff Sleeve", "Ruched", "Satin", "Sequins", "Shimmer", "Short", "Short Sleeve", "Side Slit", "Square Neck", "Strapless", "Sweetheart Neck", "Tight", "V-Neck", "Velvet", "Wrap"

How to use

import torch
import torchvision.transforms as transforms
from PIL import Image
from safetensors.torch import load_model
from huggingface_hub import hf_hub_download
from timm import list_models, create_model
import os
import numpy as np

# Download model from hub
os.makedirs('/content/swin_s3_base_224', exist_ok=True)
hf_hub_download(repo_id="LucyintheSky/lucy-feature-prediction", filename="model.safetensors", local_dir="/content/swin_s3_base_224")

# Intialize the model
model_name='swin_s3_base_224'
model = create_model(
    model_name,
    num_classes=36
)
load_model(model,f'./{model_name}/model.safetensors')

# Define class names
class_names = ["3/4 Sleeve", "Accessory", "Babydoll", "Closed Back", "Corset", "Crochet", "Cutouts", "Draped", "Floral", "Gloves", "Halter", "Lace", "Long", "Long Sleeve", "Midi", "No Slit", "Off The Shoulder", "One Shoulder", "Open Back", "Pockets", "Print", "Puff Sleeve", "Ruched", "Satin", "Sequins", "Shimmer", "Short", "Short Sleeve", "Side Slit", "Square Neck", "Strapless", "Sweetheart Neck", "Tight", "V-Neck", "Velvet", "Wrap"]
label2id = {c:idx for idx,c in enumerate(class_names)}
id2label = {idx:c for idx,c in enumerate(class_names)}

def predict_features(image_path):
    # Load PIL image
    pil_image = Image.open(image_path).convert('RGB')

    # Define transformations to resize and convert image to tensor
    transform = transforms.Compose([
        transforms.Resize((224, 224)),
        transforms.ToTensor()
    ])
    tensor_image = transform(pil_image)

    inputs = tensor_image.unsqueeze(0)

    with torch.no_grad():
        logits = model(inputs)

    # apply sigmoid activation to convert logits to probabilities
    # getting labels with confidence threshold of 0.5
    predictions = logits.sigmoid() > 0.5

    # converting one-hot encoded predictions back to list of labels
    predictions = predictions.float().numpy().flatten() # convert boolean predictions to float
    pred_labels = np.where(predictions==1)[0] # find indices where prediction is 1
    pred_labels = ([id2label[label] for label in pred_labels]) # converting integer labels to string

    return pred_labels

print(predict_features('image.jpg'))

References

https://huggingface.co/learn/computer-vision-course/en/unit3/vision-transformers/vision-transformers-for-image-classification