app.py

#app.py:
# from huggingface_hub import from_pretrained_fastai
import gradio as gr

from fastai import *
from fastai.data.block import DataBlock
from fastai.data.transforms import get_image_files, FuncSplitter, Normalize
from fastai.layers import Mish
from fastai.losses import BaseLoss
from fastai.optimizer import ranger
from fastai.torch_core import tensor
from fastai.vision.augment import aug_transforms
from fastai.vision.core import PILImage, PILMask
from fastai.vision.data import ImageBlock, MaskBlock, imagenet_stats
from fastai.vision.learner import unet_learner
from PIL import Image
import numpy as np
from torch import nn
import torch
import torch.nn.functional as F


#from __future__ import annotations
#from nbdev.showdoc import *
#from fastai import fastcore
from fastcore.test import *
from fastcore.nb_imports import *

from fastcore.imports import *
from fastcore.foundation import *
from fastcore.utils import *
from fastcore.dispatch import *
from fastcore.transform import *
import inspect

# # repo_id = "YOUR_USERNAME/YOUR_LEARNER_NAME"
# repo_id = "islasher/segm-grapes"
# repo_id='islasher/segm-grapes'


# # Definimos una función que se encarga de llevar a cabo las predicciones

# from fastai.learner import load_learner

# # Cargar el modelo y el tokenizador
# learn = load_learner(repo_id)
#learner = from_pretrained_fastai(repo_id)


class ItemTransform(Transform):
    "A transform that always take tuples as items"
    _retain = True
    def __call__(self, x, **kwargs): return self._call1(x, '__call__', **kwargs)
    def decode(self, x, **kwargs):   return self._call1(x, 'decode', **kwargs)
    def _call1(self, x, name, **kwargs):
        if not _is_tuple(x): return getattr(super(), name)(x, **kwargs)
        y = getattr(super(), name)(list(x), **kwargs)
        if not self._retain: return y
        if is_listy(y) and not isinstance(y, tuple): y = tuple(y)
        return retain_type(y, x)

from huggingface_hub import from_pretrained_fastai
import torchvision.transforms as transforms
# from Transform import ItemTransform


from albumentations import (
    Compose,
    OneOf,
    ElasticTransform,
    GridDistortion,
    OpticalDistortion,
    HorizontalFlip,
    Rotate,
    Transpose,
    CLAHE,
    ShiftScaleRotate
)

class SegmentationAlbumentationsTransform(ItemTransform):
    split_idx = 0

    def __init__(self, aug):
        self.aug = aug

    def encodes(self, x):
        img,mask = x
        aug = self.aug(image=np.array(img), mask=np.array(mask))
        return PILImage.create(aug["image"]), PILMask.create(aug["mask"])

class TargetMaskConvertTransform(ItemTransform):
    def __init__(self):
        pass
    def encodes(self, x):
        img,mask = x

        #Convert to array
        mask = np.array(mask)


        # Changes: (codes= array(['Background', 'Leaves', 'Wood', 'Pole', 'Grape'], dtype='<U10'))
        mask[mask==150]=1 #leaves

        mask[mask==76]=3 #pole
        mask[mask==74]=3 #pole

        mask[mask==29]=2 #wood
        mask[mask==25]=2 #wood

        mask[mask==255]=4 #grape

        mask[mask==0]=0
        # Back to PILMask
        mask = PILMask.create(mask)
        return img, mask



def get_y_fn (x):
    return Path(str(x).replace("Images","Labels").replace("color","gt").replace(".jpg",".png"))
    
learn = from_pretrained_fastai("islasher/segm-grapes")
#device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
def transform_image(image):
    my_transforms = transforms.Compose([transforms.ToTensor(),
                                        transforms.Normalize(
                                            [0.485, 0.456, 0.406],
                                            [0.229, 0.224, 0.225])])
    image_aux = image
    return my_transforms(image_aux).unsqueeze(0)#.to(device)






# Definimos una función que se encarga de llevar a cabo las predicciones
def predict(img):
    img=Image.fromarray(img)
    image = transforms.Resize((480,640))(img)
    tensor = transform_image(image=image)
    with torch.no_grad():
        outputs = learn.model(tensor)
    
    outputs = torch.argmax(outputs,1)

    mask = np.array(outputs)
    mask[mask==1]=150
    mask[mask==3]=76 #pole # y no 74
# mask[mask==5]=74 #pole
    mask[mask==2]=29 #wood # y no 25
# mask[mask==6]=25 #wood
    mask[mask==4]=255 #grape
    mask=np.reshape(mask,(480,640)) #en modo matriz
    return Image.fromarray(mask.astype('uint8'))
    
# Creamos la interfaz y la lanzamos. 
gr.Interface(fn=predict, inputs=gr.Image(), outputs=gr.Image(),examples=['color_154.jpg','color_155.jpg']).launch(share=False) #shape=(128, 128)  shape=(480,640)