Spaces:
Sleeping
Sleeping
File size: 1,025 Bytes
b1a427a 31fd1a2 b1a427a 76b3196 b1a427a 80c2f4a b1a427a a09bc7c b1a427a a09bc7c b1a427a 1be2d5d b1a427a |
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 |
from PIL import Image
import requests
import gradio as gr
import torch
from loader import get_loader
import torchvision.transforms as transforms
transform = transforms.Compose([
transforms.Resize(256),
transforms.CenterCrop(224),
transforms.ToTensor(),
transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]),
])
train_loader,dataset=get_loader(root_folder='FlickrDataset/Images',annotation_file='captions.txt',transform=transform,num_workers=2)
filepath="ImageCaptioningusingLSTM.pth"
from model import CNNtoRNN
model=CNNtoRNN(embed_size=256,hidden_size=256,vocab_size=len(dataset.vocab),num_layers=1)
model.load_state_dict(torch.load(filepath,map_location='cpu'))
model.eval()
def launch(input):
im=Image.open(requests.get(input,stream=True).raw)
image=transform(im.convert('RGB')).unsqueeze(0)
output=model.caption_image(image,dataset.vocab)
str=" ".join(output[1:-1])
return str
iface=gr.Interface(launch,inputs="text",outputs="text")
iface.launch()
|