nprasad2021

import sys

SCORES = [7, 15, 12, 44, 12]
OPTIONS = [2, 3, 3, 1, 5]
ISSUES = ["A", "B", "C", "D", "E"]

NAMES = {
	"A": "Data Sharing",
	"B": "Seeds and Capital",
	"C": "Restrictions on Local Genetic Resources",

nprasad2021

import numpy as np
import tensorflow as tf
from rdkit import Chem
from rdkit.Chem import AllChem
from tensorflow.keras.models import Sequential
from abc import ABC, abstractmethod

MODEL_PATHS = {
    'n450': '/home/sandia/OLED_HTL_PRIME/CPL_ver/training/N2_750/models/n450.h5',
    'n530': '/home/sandia/OLED_HTL_PRIME/CPL_ver/training/N2_750/models/n530.h5',

nprasad2021

How to train and test the head detector

Neeraj Prasad

Setup and Installation

Run the following instructions in bash in your Linux Environment with GPUs

git clone git@gist.github.com:9c3d406f69580929fe86023154dd4390.git
mv 9c3d406f69580929fe86023154dd4390/head_detector.sh ./head_detector.sh
chmod +x head_detector.sh

nprasad2021

#!/bin/bash
mkdir head_detector
cd head_detector
git clone git@github.com:nprasad2021/maskrcnn-benchmark.git
cd maskrcnn-benchmark
python3 -m venv ./mask
source ./mask/bin/activate
pip3 install ninja yacs cython matplotlib xmltodict future
pip3 install opencv-python==3.4.5.20 sklearn requests
pip3 install torch torchvision

nprasad2021

with tf.Session() as sess:
	# Instantiate all variables
	sess.run(global_init)

	training_handle = sess.run(train_iterator.string_handle())
	validation_handle = sess.run(val_iterator.string_handle())


	num_epochs = 20

nprasad2021

total_loss = tf.losses.get_total_loss()
global_step = tf.Variable(0, trainable=False)

params = tf.trainable_variables()
gradients = tf.gradients(total_loss, params)

optimizer = tf.train.AdamOptimizer(learning_rate=0.001)
updates = optimizer.apply_gradients(zip(gradients, params), global_step=global_step)
global_init = tf.variables_initializer(tf.global_variables())

nprasad2021

logits = inference(left_input_im, right_input_im)
loss(logits, left_label, right_label)

nprasad2021

train_dataset = combine_dataset(batch_size=128, image_size=224, same_prob=0.95, diff_prob=.001, train=True)
val_dataset = combine_dataset(batch_size=128, image_size=224, same_prob=0.95, diff_prob=.001, train=False)
handle = tf.placeholder(tf.string, shape=[])

iterator = tf.data.Iterator.from_string_handle(
            handle, train_dataset.output_types, train_dataset.output_shapes)

train_iterator = train_dataset.make_one_shot_iterator()
val_iterator = val_dataset.make_one_shot_iterator()

nprasad2021

def loss(logits, left_label, right_label):
	label = tf.equal(left_label, right_label)
	label = tf.cast(label, tf.int32)

	cross_entropy_loss = tf.reduce_mean(tf.nn.sigmoid_cross_entropy_with_logits(logits=logits, labels=label))
	tf.losses.add_loss(cross_entropy_loss)

nprasad2021

def inference(left_input_image, right_input_image)
	'''
	left_input_image: 3D tensor input
	right_input_image: 3D tensor input
	label: 1 if images are from same category. 0 if not.
	'''
	with tf.variable_scope('feature_generator', reuse=tf.AUTO_REUSE) as sc:

 left_features = model(left_input_image)