I have a little bit trouble with deploying streamlit app and loading keras models.
If i run it locally it works fine.
AttributeError: 'str' object has no attribute 'decode'
Traceback:
File "/usr/local/lib/python3.7/site-packages/streamlit/script_runner.py", line 332, in _run_script
exec(code, module.__dict__)File "/app/app_testing/streamlit_app.py", line 74, in <module>
model_eval, model_auto = load_models()File "/usr/local/lib/python3.7/site-packages/streamlit/caching.py", line 591, in wrapped_func
return get_or_create_cached_value()File "/usr/local/lib/python3.7/site-packages/streamlit/caching.py", line 575, in get_or_create_cached_value
return_value = func(*args, **kwargs)File "/app/app_testing/streamlit_app.py", line 20, in load_models
model_eval = load_model('models/doc_model.h5')File "/home/appuser/.local/lib/python3.7/site-packages/tensorflow_core/python/keras/saving/save.py", line 146, in load_model
return hdf5_format.load_model_from_hdf5(filepath, custom_objects, compile)File "/home/appuser/.local/lib/python3.7/site-packages/tensorflow_core/python/keras/saving/hdf5_format.py", line 166, in load_model_from_hdf5
model_config = json.loads(model_config.decode('utf-8'))
import streamlit as st
import io
import cv2
from PIL import Image
import textract
import numpy as np
import tensorflow as tf
from tensorflow.keras.models import load_model
from tensorflow.keras.preprocessing.image import img_to_array, load_img, array_to_img
CLASS_IDXS = ["not good", "good"]
@st.cache(allow_output_mutation=True)
def load_models():
model_eval = load_model('models/doc_model.h5')
model_eval._make_predict_function()
model_eval.summary()
model_auto = load_model('models/auto_model.h5')
model_auto._make_predict_function()
model_auto.summary()
return model_eval, model_auto
@st.cache
def __calculate_score(y_pred_class, y_pred_prob):
if y_pred_class == 0:
MAX = 0.5
scaled_percentage = (y_pred_prob * MAX) / 100
return MAX - scaled_percentage
else:
MAX = 1
scaled_percentage = (y_pred_prob * MAX) / 100
return scaled_percentage
@st.cache
def __load_and_preprocess_custom_image(image_path):
img = load_img(image_path, color_mode = 'grayscale', target_size = (700, 700))
img = img_to_array(img).astype('float32')/255
return img
@st.cache
def __predict_score(image):
image = __load_and_preprocess_custom_image(image)
y_pred = model_eval.predict(np.expand_dims(image, axis=0), verbose=1)[0]
y_pred_class = np.argmax(y_pred)
y_pred_prob = y_pred[y_pred_class]*100
score = __calculate_score(y_pred_class, y_pred_prob)
return y_pred_class, score
@st.cache
def __auto_encode(image):
org_img = load_img(image, color_mode = 'grayscale')
org_img = img_to_array(org_img)
img = org_img.astype('float32')
img = np.expand_dims(img, axis=0)
y_pred = np.squeeze(model_auto.predict(img, verbose=1))
img = cv2.convertScaleAbs(y_pred, alpha=(255.0))
img = Image.fromarray(img)
return img
@st.cache
def __get_text_from_image(image):
text = textract.process(image, method='tesseract', encoding='utf-8')
text = text.decode('utf8')
return text
st.title("denoise and evaluate images")
img_file_buffer = st.file_uploader("Upload an image", type=["png", "jpg", "jpeg"])
model_eval, model_auto = load_models()
if img_file_buffer is not None:
org = load_img(img_file_buffer)
y_pred_class, score = __predict_score(img_file_buffer)
text = __get_text_from_image(img_file_buffer)
st.image(org, caption=f"Original", width=700)
st.write("Predicted class : %s" % (CLASS_IDXS[y_pred_class]))
st.write("Score : %f" % (score))
st.write(text)
img = __auto_encode(img_file_buffer)
file_object = io.BytesIO()
img.save(file_object, 'PNG')
y_pred_class, score = __predict_score(file_object)
text = __get_text_from_image(file_object)
st.image(img, caption=f"Processed Image", width=700)
st.write("Predicted class : %s" % (CLASS_IDXS[y_pred_class]))
st.write("Score : %f" % (score))
st.write(text)
else:
st.write('Please upload single image')
