Spaces:
Sleeping
Sleeping
File size: 3,552 Bytes
19c08c2 efb91fe 19c08c2 572aec8 efb91fe 19c08c2 efb91fe 19c08c2 6547b74 18940a0 31f7642 19c08c2 31f7642 19c08c2 efb91fe 19c08c2 6547b74 |
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 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 |
import streamlit as st
import pickle
import tensorflow as tf
import cv2
import numpy as np
from PIL import Image, ImageOps
import imageio.v3 as iio
import time
from textwrap import wrap
import matplotlib.pylab as plt
from tensorflow.keras import Input
from tensorflow.keras.layers import (
GRU,
Add,
AdditiveAttention,
Attention,
Concatenate,
Dense,
Embedding,
LayerNormalization,
Reshape,
StringLookup,
TextVectorization,
)
@st.cache_resource()
def load_image_model():
image_model=tf.keras.models.load_model('./image_caption_model.h5')
return image_model
@st.cache_resource()
def load_decoder_model():
decoder_model=tf.keras.models.load_model('./decoder_pred_model.h5')
return decoder_model
@st.cache_resource()
def load_encoder_model():
encoder=tf.keras.models.load_model('./encoder_model.h5')
return encoder
st.title(":blue[Nishant Guvvada's] :red[AI Journey] Image Caption Generation")
image = Image.open('./title.jpg')
st.image(image)
st.write("""
# Multi-Modal Machine Learning
"""
)
file = st.file_uploader("Upload any image and the model will try to provide a caption to it!", type= ['png', 'jpg'])
MAX_CAPTION_LEN = 64
MINIMUM_SENTENCE_LENGTH = 5
IMG_HEIGHT = 299
IMG_WIDTH = 299
IMG_CHANNELS = 3
ATTENTION_DIM = 512 # size of dense layer in Attention
VOCAB_SIZE = 20000
# We will override the default standardization of TextVectorization to preserve
# "<>" characters, so we preserve the tokens for the <start> and <end>.
def standardize(inputs):
inputs = tf.strings.lower(inputs)
return tf.strings.regex_replace(
inputs, r"[!\"#$%&\(\)\*\+.,-/:;=?@\[\\\]^_`{|}~]?", ""
)
# Choose the most frequent words from the vocabulary & remove punctuation etc.
vocab = open('./tokenizer_vocab.txt', 'rb')
tokenizer = pickle.load(vocab)
# Lookup table: Word -> Index
word_to_index = StringLookup(
mask_token="", vocabulary=tokenizer
)
## Probabilistic prediction using the trained model
def predict_caption(file):
filename = Image.open(file)
image = filename.convert('RGB')
image = np.array(image)
gru_state = tf.zeros((1, ATTENTION_DIM))
resize = tf.image.resize(image, (IMG_HEIGHT, IMG_WIDTH))
img = np.expand_dims(resize/255, 0)
encoder = load_encoder_model()
features = encoder(tf.expand_dims(img, axis=0))
dec_input = tf.expand_dims([word_to_index("<start>")], 1)
result = []
decoder_pred_model = load_decoder_model()
for i in range(MAX_CAPTION_LEN):
predictions, gru_state = decoder_pred_model(
[dec_input, gru_state, features]
)
# draws from log distribution given by predictions
top_probs, top_idxs = tf.math.top_k(
input=predictions[0][0], k=10, sorted=False
)
chosen_id = tf.random.categorical([top_probs], 1)[0].numpy()
predicted_id = top_idxs.numpy()[chosen_id][0]
result.append(tokenizer[predicted_id])
if predicted_id == word_to_index("<end>"):
return img, result
dec_input = tf.expand_dims([predicted_id], 1)
return img, result
def on_click():
if file is None:
st.text("Please upload an image file")
else:
image = Image.open(file)
st.image(image, use_column_width=True)
for i in range(5):
image, caption = predict_caption(file)
#print(" ".join(caption[:-1]) + ".")
st.write(" ".join(caption[:-1]) + ".")
st.button('Generate', on_click=on_click) |