Скачать или смотреть 31. Deploy Model pada Web Server dengan FastAPI

The codes can be downloaded from:
https://github.com/nemuelpah/Code-for...

The model myNewModel.h5 is not included in github due to its size. You can create your own model using the steps in Video 29.

If you have a problem to load h5 file. Save and load using JSON as in
https://machinelearningmastery.com/sa...

You need to have the saved files (model.h5 and model.json) in the folder. So your proses.py file should be like the following:

from PIL import Image
from keras.applications.mobilenet import preprocess_input
import numpy as np
import keras
from keras.models import load_model
from numpy import asarray
from tensorflow.keras.models import Sequential, model_from_json

def proses(file):
#model_baru=load_model('myNewModel.h5')
json_file = open('model.json', 'r')
loaded_model_json = json_file.read()
json_file.close()
model_baru = model_from_json(loaded_model_json)
load weights into new model
model_baru.load_weights("model.h5")

jenis = ['Parang','Mega Mendung','Kawung']

image = Image.open(file.file)
image = image.resize((160,160))
image = np.asarray(image)
image = np.expand_dims(image,0)

p = model_baru.predict(image)
kelas = p.argmax(axis = 1)[0]
label = jenis[kelas]
conf = p[0][kelas]
return conf, label