icing.py 31.11 KiB
import tensorflow as tf
import tensorflow_addons as tfa
from util.setup import logdir, modeldir, cachepath, now
from util.util import homedir, EarlyStop
from util.geos_nav import GEOSNavigation
import os, datetime
import numpy as np
import pickle
import h5py
from icing.pirep_goes import normalize, make_for_full_domain_predict
LOG_DEVICE_PLACEMENT = False
CACHE_DATA_IN_MEM = False
PROC_BATCH_SIZE = 4096
PROC_BATCH_BUFFER_SIZE = 50000
NumClasses = 2
if NumClasses == 2:
NumLogits = 1
else:
NumLogits = NumClasses
BATCH_SIZE = 128
NUM_EPOCHS = 100
TRACK_MOVING_AVERAGE = False
EARLY_STOP = False
TRIPLET = False
CONV3D = False
NOISE_TRAINING = False
img_width = 16
mean_std_file = homedir+'data/icing/mean_std_no_ice.pkl'
# mean_std_file = homedir+'data/icing/mean_std_l1b_no_ice.pkl'
f = open(mean_std_file, 'rb')
mean_std_dct = pickle.load(f)
f.close()
# train_params = ['cld_height_acha', 'cld_geo_thick', 'supercooled_cloud_fraction', 'cld_temp_acha', 'cld_press_acha',
# 'cld_reff_acha', 'cld_opd_acha', 'conv_cloud_fraction', 'cld_emiss_acha']
train_params = ['cld_height_acha', 'cld_geo_thick', 'supercooled_cloud_fraction', 'cld_temp_acha', 'cld_press_acha',
'cld_reff_dcomp', 'cld_opd_dcomp', 'cld_cwp_dcomp', 'iwc_dcomp', 'lwc_dcomp', 'conv_cloud_fraction', 'cld_emiss_acha']
# train_params = ['temp_10_4um_nom', 'temp_11_0um_nom', 'temp_12_0um_nom', 'temp_13_3um_nom', 'temp_3_75um_nom',
# 'temp_6_2um_nom', 'temp_6_7um_nom', 'temp_7_3um_nom', 'temp_8_5um_nom', 'temp_9_7um_nom',
# 'refl_0_47um_nom', 'refl_0_65um_nom', 'refl_0_86um_nom', 'refl_1_38um_nom', 'refl_1_60um_nom']
# train_params = ['temp_10_4um_nom', 'temp_11_0um_nom', 'temp_12_0um_nom', 'temp_13_3um_nom', 'temp_3_75um_nom',
# 'temp_6_2um_nom', 'temp_6_7um_nom', 'temp_7_3um_nom', 'temp_8_5um_nom', 'temp_9_7um_nom']
def build_residual_block(input, drop_rate, num_neurons, activation, block_name, doDropout=True, doBatchNorm=True):
with tf.name_scope(block_name):
if doDropout:
fc = tf.keras.layers.Dropout(drop_rate)(input)
fc = tf.keras.layers.Dense(num_neurons, activation=activation)(fc)
else:
fc = tf.keras.layers.Dense(num_neurons, activation=activation)(input)
if doBatchNorm:
fc = tf.keras.layers.BatchNormalization()(fc)
print(fc.shape)
fc_skip = fc
if doDropout:
fc = tf.keras.layers.Dropout(drop_rate)(fc)
fc = tf.keras.layers.Dense(num_neurons, activation=activation)(fc)
if doBatchNorm:
fc = tf.keras.layers.BatchNormalization()(fc)
print(fc.shape)
if doDropout:
fc = tf.keras.layers.Dropout(drop_rate)(fc)
fc = tf.keras.layers.Dense(num_neurons, activation=activation)(fc)
if doBatchNorm:
fc = tf.keras.layers.BatchNormalization()(fc)
print(fc.shape)
if doDropout:
fc = tf.keras.layers.Dropout(drop_rate)(fc)
fc = tf.keras.layers.Dense(num_neurons, activation=None)(fc)
if doBatchNorm:
fc = tf.keras.layers.BatchNormalization()(fc)
fc = fc + fc_skip
fc = tf.keras.layers.LeakyReLU()(fc)
print(fc.shape)
return fc
class IcingIntensityNN:
def __init__(self, gpu_device=0, datapath=None):
self.train_data = None
self.train_label = None
self.test_data = None
self.test_label = None
self.test_data_denorm = None
self.train_dataset = None
self.inner_train_dataset = None
self.test_dataset = None
self.eval_dataset = None
self.X_img = None
self.X_prof = None
self.X_u = None
self.X_v = None
self.X_sfc = None
self.inputs = []
self.y = None
self.handle = None
self.inner_handle = None
self.in_mem_batch = None
self.filename_trn = None
self.h5f_trn = None
self.filename_tst = None
self.h5f_tst = None
self.h5f_l1b = None
self.logits = None
self.predict_data = None
self.predict_dataset = None
self.mean_list = None
self.std_list = None
self.training_op = None
self.correct = None
self.accuracy = None
self.loss = None
self.pred_class = None
self.gpu_device = gpu_device
self.variable_averages = None
self.global_step = None
self.writer_train = None
self.writer_valid = None
self.OUT_OF_RANGE = False
self.abi = None
self.temp = None
self.wv = None
self.lbfp = None
self.sfc = None
self.in_mem_data_cache = {}
self.model = None
self.optimizer = None
self.train_loss = None
self.train_accuracy = None
self.test_loss = None
self.test_accuracy = None
self.test_auc = None
self.test_recall = None
self.test_precision = None
self.test_confusion_matrix = None
self.test_true_pos = None
self.test_true_neg = None
self.test_false_pos = None
self.test_false_neg = None
self.test_labels = []
self.test_preds = []
self.test_probs = None
self.learningRateSchedule = None
self.num_data_samples = None
self.initial_learning_rate = None
self.data_dct = None
n_chans = len(train_params)
if TRIPLET:
n_chans *= 3
self.X_img = tf.keras.Input(shape=n_chans)
self.inputs.append(self.X_img)
self.DISK_CACHE = False
if datapath is not None:
self.DISK_CACHE = False
f = open(datapath, 'rb')
self.in_mem_data_cache = pickle.load(f)
f.close()
tf.debugging.set_log_device_placement(LOG_DEVICE_PLACEMENT)
gpus = tf.config.experimental.list_physical_devices('GPU')
if gpus:
try:
# Currently, memory growth needs to be the same across GPUs
for gpu in gpus:
tf.config.experimental.set_memory_growth(gpu, True)
logical_gpus = tf.config.experimental.list_logical_devices('GPU')
print(len(gpus), "Physical GPUs,", len(logical_gpus), "Logical GPUs")
except RuntimeError as e:
# Memory growth must be set before GPUs have been initialized
print(e)
def get_in_mem_data_batch(self, idxs, is_training):
h5f = self.h5f_trn
if not is_training:
h5f = self.h5f_tst
key = frozenset(idxs)
if CACHE_DATA_IN_MEM:
tup = self.in_mem_data_cache.get(key)
if tup is not None:
return tup[0], tup[1]
# sort these to use as numpy indexing arrays
nd_idxs = np.array(idxs)
nd_idxs = np.sort(nd_idxs)
data = []
for param in train_params:
nda = h5f[param][nd_idxs, ]
if NOISE_TRAINING and is_training:
nda = normalize(nda, param, mean_std_dct, add_noise=True, noise_scale=0.01, seed=42)
else:
nda = normalize(nda, param, mean_std_dct)
data.append(nda)
data = np.stack(data)
data = data.astype(np.float32)
data = np.transpose(data, axes=(1, 0))
label = h5f['icing_intensity'][nd_idxs]
label = label.astype(np.int32)
label = np.where(label == -1, 0, label)
# binary, two class
if NumClasses == 2:
label = np.where(label != 0, 1, label)
label = label.reshape((label.shape[0], 1))
elif NumClasses == 3:
label = np.where(np.logical_or(label == 1, label == 2), 1, label)
label = np.where(np.invert(np.logical_or(label == 0, label == 1)), 2, label)
label = label.reshape((label.shape[0], 1))
if CACHE_DATA_IN_MEM:
self.in_mem_data_cache[key] = (data, label)
return data, label
def get_in_mem_data_batch_train(self, idxs):
return self.get_in_mem_data_batch(idxs, True)
def get_in_mem_data_batch_test(self, idxs):
return self.get_in_mem_data_batch(idxs, False)
def get_in_mem_data_batch_eval(self, idxs):
# sort these to use as numpy indexing arrays
nd_idxs = np.array(idxs)
nd_idxs = np.sort(nd_idxs)
data = []
for param in train_params:
nda = self.data_dct[param][nd_idxs, ]
nda = normalize(nda, param, mean_std_dct)
data.append(nda)
data = np.stack(data)
data = data.astype(np.float32)
data = np.transpose(data, axes=(1,0))
return data
@tf.function(input_signature=[tf.TensorSpec(None, tf.int32)])
def data_function(self, indexes):
out = tf.numpy_function(self.get_in_mem_data_batch_train, [indexes], [tf.float32, tf.int32])
return out
@tf.function(input_signature=[tf.TensorSpec(None, tf.int32)])
def data_function_test(self, indexes):
out = tf.numpy_function(self.get_in_mem_data_batch_test, [indexes], [tf.float32, tf.int32])
return out
@tf.function(input_signature=[tf.TensorSpec(None, tf.int32)])
def data_function_evaluate(self, indexes):
out = tf.numpy_function(self.get_in_mem_data_batch_eval, [indexes], tf.float32)
return out
def get_train_dataset(self, indexes):
indexes = list(indexes)
dataset = tf.data.Dataset.from_tensor_slices(indexes)
dataset = dataset.batch(PROC_BATCH_SIZE)
dataset = dataset.map(self.data_function, num_parallel_calls=8)
dataset = dataset.cache()
# dataset = dataset.shuffle(PROC_BATCH_BUFFER_SIZE)
dataset = dataset.prefetch(buffer_size=1)
self.train_dataset = dataset
def get_test_dataset(self, indexes):
indexes = list(indexes)
dataset = tf.data.Dataset.from_tensor_slices(indexes)
dataset = dataset.batch(PROC_BATCH_SIZE)
dataset = dataset.map(self.data_function_test, num_parallel_calls=8)
dataset = dataset.cache()
self.test_dataset = dataset
def get_evaluate_dataset(self, indexes):
indexes = list(indexes)
dataset = tf.data.Dataset.from_tensor_slices(indexes)
dataset = dataset.batch(PROC_BATCH_SIZE)
dataset = dataset.map(self.data_function_evaluate, num_parallel_calls=8)
dataset = dataset.cache()
self.eval_dataset = dataset
def setup_pipeline(self, filename_trn, filename_tst, trn_idxs=None, tst_idxs=None, seed=None):
self.filename_trn = filename_trn
self.h5f_trn = h5py.File(filename_trn, 'r')
self.filename_tst = filename_tst
self.h5f_tst = h5py.File(filename_tst, 'r')
if trn_idxs is None:
time = self.h5f_trn['time']
trn_idxs = np.arange(time.shape[0])
if seed is not None:
np.random.seed(seed)
np.random.shuffle(trn_idxs)
time = self.h5f_tst['time']
tst_idxs = np.arange(time.shape[0])
if seed is not None:
np.random.seed(seed)
np.random.shuffle(tst_idxs)
self.num_data_samples = trn_idxs.shape[0]
self.get_train_dataset(trn_idxs)
self.get_test_dataset(tst_idxs)
print('num train samples: ', self.num_data_samples)
print('BATCH SIZE: ', BATCH_SIZE)
print('num test samples: ', tst_idxs.shape[0])
print('setup_pipeline: Done')
def setup_test_pipeline(self, filename, seed=None, shuffle=False):
self.filename_tst = filename
self.h5f_tst = h5py.File(filename, 'r')
time = self.h5f_tst['time']
tst_idxs = np.arange(time.shape[0])
self.num_data_samples = len(tst_idxs)
if seed is not None:
np.random.seed(seed)
if shuffle:
np.random.shuffle(tst_idxs)
self.get_test_dataset(tst_idxs)
print('num test samples: ', tst_idxs.shape[0])
print('setup_test_pipeline: Done')
def setup_eval_pipeline(self, data_dct, num_tiles):
self.data_dct = data_dct
idxs = np.arange(num_tiles)
self.num_data_samples = idxs.shape[0]
self.get_evaluate_dataset(idxs)
def build_1d_cnn(self):
print('build_1d_cnn')
# padding = 'VALID'
padding = 'SAME'
# activation = tf.nn.relu
# activation = tf.nn.elu
activation = tf.nn.leaky_relu
num_filters = 6
conv = tf.keras.layers.Conv1D(num_filters, 5, strides=1, padding=padding)(self.inputs[1])
conv = tf.keras.layers.MaxPool1D(padding=padding)(conv)
print(conv)
num_filters *= 2
conv = tf.keras.layers.Conv1D(num_filters, 3, strides=1, padding=padding)(conv)
conv = tf.keras.layers.MaxPool1D(padding=padding)(conv)
print(conv)
num_filters *= 2
conv = tf.keras.layers.Conv1D(num_filters, 3, strides=1, padding=padding)(conv)
conv = tf.keras.layers.MaxPool1D(padding=padding)(conv)
print(conv)
num_filters *= 2
conv = tf.keras.layers.Conv1D(num_filters, 3, strides=1, padding=padding)(conv)
conv = tf.keras.layers.MaxPool1D(padding=padding)(conv)
print(conv)
flat = tf.keras.layers.Flatten()(conv)
print(flat)
return flat
def build_dnn(self, input_layer=None):
print('build fully connected layer')
drop_rate = 0.5
# activation = tf.nn.relu
# activation = tf.nn.elu
activation = tf.nn.leaky_relu
momentum = 0.99
if input_layer is not None:
flat = input_layer
n_hidden = input_layer.shape[1]
else:
flat = self.X_img
n_hidden = self.X_img.shape[1]
fac = 2
fc = build_residual_block(flat, drop_rate, fac*n_hidden, activation, 'Residual_Block_1', doBatchNorm=True)
fc = build_residual_block(fc, drop_rate, fac*n_hidden, activation, 'Residual_Block_2', doBatchNorm=True)
fc = build_residual_block(fc, drop_rate, fac*n_hidden, activation, 'Residual_Block_3', doBatchNorm=True)
fc = build_residual_block(fc, drop_rate, fac*n_hidden, activation, 'Residual_Block_4', doBatchNorm=True)
fc = build_residual_block(fc, drop_rate, fac*n_hidden, activation, 'Residual_Block_5', doBatchNorm=True)
fc = build_residual_block(fc, drop_rate, fac*n_hidden, activation, 'Residual_Block_6', doBatchNorm=True)
fc = build_residual_block(fc, drop_rate, fac*n_hidden, activation, 'Residual_Block_7', doBatchNorm=True)
fc = build_residual_block(fc, drop_rate, fac*n_hidden, activation, 'Residual_Block_8', doBatchNorm=True)
fc = build_residual_block(fc, drop_rate, fac*n_hidden, activation, 'Residual_Block_9', doBatchNorm=True)
fc = build_residual_block(fc, drop_rate, fac*n_hidden, activation, 'Residual_Block_10', doBatchNorm=True)
fc = tf.keras.layers.Dense(n_hidden, activation=activation)(fc)
fc = tf.keras.layers.BatchNormalization()(fc)
print(fc.shape)
if NumClasses == 2:
activation = tf.nn.sigmoid # For binary
else:
activation = tf.nn.softmax # For multi-class
# Called logits, but these are actually probabilities, see activation
logits = tf.keras.layers.Dense(NumLogits, activation=activation)(fc)
print(logits.shape)
self.logits = logits
def build_training(self):
if NumClasses == 2:
self.loss = tf.keras.losses.BinaryCrossentropy(from_logits=False) # for two-class only
else:
self.loss = tf.keras.losses.SparseCategoricalCrossentropy(from_logits=False) # For multi-class
# decayed_learning_rate = learning_rate * decay_rate ^ (global_step / decay_steps)
initial_learning_rate = 0.002
decay_rate = 0.95
steps_per_epoch = int(self.num_data_samples/BATCH_SIZE) # one epoch
# decay_steps = int(steps_per_epoch / 2)
decay_steps = 8 * steps_per_epoch
print('initial rate, decay rate, steps/epoch, decay steps: ', initial_learning_rate, decay_rate, steps_per_epoch, decay_steps)
self.learningRateSchedule = tf.keras.optimizers.schedules.ExponentialDecay(initial_learning_rate, decay_steps, decay_rate)
optimizer = tf.keras.optimizers.Adam(learning_rate=self.learningRateSchedule)
if TRACK_MOVING_AVERAGE: # Not really sure this works properly
optimizer = tfa.optimizers.MovingAverage(optimizer)
self.optimizer = optimizer
self.initial_learning_rate = initial_learning_rate
def build_evaluation(self):
self.train_loss = tf.keras.metrics.Mean(name='train_loss')
self.test_loss = tf.keras.metrics.Mean(name='test_loss')
if NumClasses == 2:
self.train_accuracy = tf.keras.metrics.BinaryAccuracy(name='train_accuracy')
self.test_accuracy = tf.keras.metrics.BinaryAccuracy(name='test_accuracy')
self.test_auc = tf.keras.metrics.AUC(name='test_auc')
self.test_recall = tf.keras.metrics.Recall(name='test_recall')
self.test_precision = tf.keras.metrics.Precision(name='test_precision')
self.test_true_neg = tf.keras.metrics.TrueNegatives(name='test_true_neg')
self.test_true_pos = tf.keras.metrics.TruePositives(name='test_true_pos')
self.test_false_neg = tf.keras.metrics.FalseNegatives(name='test_false_neg')
self.test_false_pos = tf.keras.metrics.FalsePositives(name='test_false_pos')
else:
self.train_accuracy = tf.keras.metrics.SparseCategoricalAccuracy(name='train_accuracy')
self.test_accuracy = tf.keras.metrics.SparseCategoricalAccuracy(name='test_accuracy')
@tf.function
def train_step(self, mini_batch):
inputs = [mini_batch[0]]
labels = mini_batch[1]
with tf.GradientTape() as tape:
pred = self.model(inputs, training=True)
loss = self.loss(labels, pred)
total_loss = loss
if len(self.model.losses) > 0:
reg_loss = tf.math.add_n(self.model.losses)
total_loss = loss + reg_loss
gradients = tape.gradient(total_loss, self.model.trainable_variables)
self.optimizer.apply_gradients(zip(gradients, self.model.trainable_variables))
self.train_loss(loss)
self.train_accuracy(labels, pred)
return loss
@tf.function
def test_step(self, mini_batch):
inputs = [mini_batch[0]]
labels = mini_batch[1]
pred = self.model(inputs, training=False)
t_loss = self.loss(labels, pred)
self.test_loss(t_loss)
self.test_accuracy(labels, pred)
if NumClasses == 2:
self.test_auc(labels, pred)
self.test_recall(labels, pred)
self.test_precision(labels, pred)
self.test_true_neg(labels, pred)
self.test_true_pos(labels, pred)
self.test_false_neg(labels, pred)
self.test_false_pos(labels, pred)
def predict(self, mini_batch):
inputs = [mini_batch[0]]
labels = mini_batch[1]
pred = self.model(inputs, training=False)
t_loss = self.loss(labels, pred)
self.test_labels.append(labels)
self.test_preds.append(pred.numpy())
self.test_loss(t_loss)
self.test_accuracy(labels, pred)
if NumClasses == 2:
self.test_auc(labels, pred)
self.test_recall(labels, pred)
self.test_precision(labels, pred)
self.test_true_neg(labels, pred)
self.test_true_pos(labels, pred)
self.test_false_neg(labels, pred)
self.test_false_pos(labels, pred)
def reset_test_metrics(self):
self.test_loss.reset_states()
self.test_accuracy.reset_states()
if NumClasses == 2:
self.test_auc.reset_states()
self.test_recall.reset_states()
self.test_precision.reset_states()
self.test_true_neg.reset_states()
self.test_true_pos.reset_states()
self.test_false_neg.reset_states()
self.test_false_pos.reset_states()
def get_metrics(self):
recall = self.test_recall.result()
precsn = self.test_precision.result()
f1 = 2 * (precsn * recall) / (precsn + recall)
tn = self.test_true_neg.result()
tp = self.test_true_pos.result()
fn = self.test_false_neg.result()
fp = self.test_false_pos.result()
mcc = ((tp * tn) - (fp * fn)) / np.sqrt((tp + fp) * (tp + fn) * (tn + fp) * (tn + fn))
return f1, mcc
def do_training(self, ckpt_dir=None):
if ckpt_dir is None:
if not os.path.exists(modeldir):
os.mkdir(modeldir)
ckpt = tf.train.Checkpoint(step=tf.Variable(1), model=self.model)
ckpt_manager = tf.train.CheckpointManager(ckpt, modeldir, max_to_keep=3)
else:
ckpt = tf.train.Checkpoint(step=tf.Variable(1), model=self.model)
ckpt_manager = tf.train.CheckpointManager(ckpt, ckpt_dir, max_to_keep=3)
self.writer_train = tf.summary.create_file_writer(os.path.join(logdir, 'plot_train'))
self.writer_valid = tf.summary.create_file_writer(os.path.join(logdir, 'plot_valid'))
step = 0
total_time = 0
best_test_loss = np.finfo(dtype=np.float).max
best_test_acc = 0
best_test_recall = 0
best_test_precision = 0
best_test_auc = 0
best_test_f1 = 0
best_test_mcc = 0
if EARLY_STOP:
es = EarlyStop()
for epoch in range(NUM_EPOCHS):
self.train_loss.reset_states()
self.train_accuracy.reset_states()
t0 = datetime.datetime.now().timestamp()
proc_batch_cnt = 0
n_samples = 0
for data0, label in self.train_dataset:
trn_ds = tf.data.Dataset.from_tensor_slices((data0, label))
trn_ds = trn_ds.batch(BATCH_SIZE)
for mini_batch in trn_ds:
if self.learningRateSchedule is not None:
loss = self.train_step(mini_batch)
if (step % 100) == 0:
with self.writer_train.as_default():
tf.summary.scalar('loss_trn', loss.numpy(), step=step)
tf.summary.scalar('num_train_steps', step, step=step)
tf.summary.scalar('num_epochs', epoch, step=step)
self.reset_test_metrics()
for data0_tst, label_tst in self.test_dataset:
tst_ds = tf.data.Dataset.from_tensor_slices((data0_tst, label_tst))
tst_ds = tst_ds.batch(BATCH_SIZE)
for mini_batch_test in tst_ds:
self.test_step(mini_batch_test)
if NumClasses == 2:
f1, mcc = self.get_metrics()
with self.writer_valid.as_default():
tf.summary.scalar('loss_val', self.test_loss.result(), step=step)
tf.summary.scalar('acc_val', self.test_accuracy.result(), step=step)
if NumClasses == 2:
tf.summary.scalar('auc_val', self.test_auc.result(), step=step)
tf.summary.scalar('recall_val', self.test_recall.result(), step=step)
tf.summary.scalar('prec_val', self.test_precision.result(), step=step)
tf.summary.scalar('f1_val', f1, step=step)
tf.summary.scalar('mcc_val', mcc, step=step)
tf.summary.scalar('num_train_steps', step, step=step)
tf.summary.scalar('num_epochs', epoch, step=step)
print('****** test loss, acc: ', self.test_loss.result().numpy(), self.test_accuracy.result().numpy())
step += 1
print('train loss: ', loss.numpy())
proc_batch_cnt += 1
n_samples += data0.shape[0]
print('proc_batch_cnt: ', proc_batch_cnt, n_samples)
t1 = datetime.datetime.now().timestamp()
print('End of Epoch: ', epoch+1, 'elapsed time: ', (t1-t0))
total_time += (t1-t0)
self.reset_test_metrics()
for data0, label in self.test_dataset:
ds = tf.data.Dataset.from_tensor_slices((data0, label))
ds = ds.batch(BATCH_SIZE)
for mini_batch in ds:
self.test_step(mini_batch)
if NumClasses == 2:
f1, mcc = self.get_metrics()
print('loss, acc, recall, precision, auc, f1, mcc: ', self.test_loss.result().numpy(), self.test_accuracy.result().numpy(),
self.test_recall.result().numpy(), self.test_precision.result().numpy(), self.test_auc.result().numpy(), f1.numpy(), mcc.numpy())
else:
print('loss, acc: ', self.test_loss.result().numpy(), self.test_accuracy.result().numpy())
print('------------------------------------------------------')
if TRACK_MOVING_AVERAGE: # This may not really work properly
self.optimizer.assign_average_vars(self.model.trainable_variables)
tst_loss = self.test_loss.result().numpy()
if tst_loss < best_test_loss:
best_test_loss = tst_loss
best_test_acc = self.test_accuracy.result().numpy()
best_test_recall = self.test_recall.result().numpy()
best_test_precision = self.test_precision.result().numpy()
best_test_auc = self.test_auc.result().numpy()
best_test_f1 = f1.numpy()
best_test_mcc = mcc.numpy()
ckpt_manager.save()
if self.DISK_CACHE and epoch == 0:
f = open(cachepath, 'wb')
pickle.dump(self.in_mem_data_cache, f)
f.close()
if EARLY_STOP and es.check_stop(tst_loss):
break
print('total time: ', total_time)
self.writer_train.close()
self.writer_valid.close()
self.h5f_trn.close()
self.h5f_tst.close()
f = open('/home/rink/best_stats_'+now+'.pkl', 'wb')
pickle.dump((best_test_loss, best_test_acc, best_test_recall, best_test_precision, best_test_auc, best_test_f1, best_test_mcc), f)
f.close()
def build_model(self):
self.build_dnn()
self.model = tf.keras.Model(self.inputs, self.logits)
def restore(self, ckpt_dir):
ckpt = tf.train.Checkpoint(step=tf.Variable(1), model=self.model)
ckpt_manager = tf.train.CheckpointManager(ckpt, ckpt_dir, max_to_keep=3)
ckpt.restore(ckpt_manager.latest_checkpoint)
self.test_loss.reset_states()
self.test_accuracy.reset_states()
for data0, label in self.test_dataset:
ds = tf.data.Dataset.from_tensor_slices((data0, label))
ds = ds.batch(BATCH_SIZE)
for mini_batch_test in ds:
self.predict(mini_batch_test)
print('loss, acc: ', self.test_loss.result(), self.test_accuracy.result())
labels = np.concatenate(self.test_labels)
self.test_labels = labels
preds = np.concatenate(self.test_preds)
self.test_probs = preds
if NumClasses == 2:
preds = np.where(preds > 0.5, 1, 0)
else:
preds = np.argmax(preds, axis=1)
self.test_preds = preds
def do_evaluate(self, ckpt_dir, prob_thresh=0.5):
ckpt = tf.train.Checkpoint(step=tf.Variable(1), model=self.model)
ckpt_manager = tf.train.CheckpointManager(ckpt, ckpt_dir, max_to_keep=3)
ckpt.restore(ckpt_manager.latest_checkpoint)
pred_s = []
for data in self.eval_dataset:
ds = tf.data.Dataset.from_tensor_slices(data)
ds = ds.batch(BATCH_SIZE)
for mini_batch in ds:
pred = self.model([mini_batch], training=False)
pred_s.append(pred)
preds = np.concatenate(pred_s)
preds = preds[:,0]
self.test_probs = preds
if NumClasses == 2:
preds = np.where(preds > prob_thresh, 1, 0)
else:
preds = np.argmax(preds, axis=1)
self.test_preds = preds
def run(self, filename_trn, filename_tst):
with tf.device('/device:GPU:'+str(self.gpu_device)):
self.setup_pipeline(filename_trn, filename_tst)
self.build_model()
self.build_training()
self.build_evaluation()
self.do_training()
def run_restore(self, filename_tst, ckpt_dir):
self.setup_test_pipeline(filename_tst)
self.build_model()
self.build_training()
self.build_evaluation()
self.restore(ckpt_dir)
self.h5f_tst.close()
def run_evaluate(self, filename, ckpt_dir):
data_dct, ll, cc = make_for_full_domain_predict(filename, name_list=train_params)
self.setup_eval_pipeline(data_dct, len(ll))
self.build_model()
self.build_training()
self.build_evaluation()
self.do_evaluate(ckpt_dir)
def run_restore_static(filename_tst, ckpt_dir_s_path):
ckpt_dir_s = os.listdir(ckpt_dir_s_path)
cm_s = []
for ckpt in ckpt_dir_s:
ckpt_dir = ckpt_dir_s_path + ckpt
if not os.path.isdir(ckpt_dir):
continue
nn = IcingIntensityNN()
nn.run_restore(filename_tst, ckpt_dir)
cm_s.append(tf.math.confusion_matrix(nn.test_labels.flatten(), nn.test_preds.flatten()))
num = len(cm_s)
cm_avg = cm_s[0]
for k in range(num-1):
cm_avg += cm_s[k+1]
cm_avg /= num
return cm_avg
def run_evaluate_static(filename, ckpt_dir_s_path, prob_thresh=0.5):
data_dct, ll, cc = make_for_full_domain_predict(filename, name_list=train_params)
ckpt_dir_s = os.listdir(ckpt_dir_s_path)
prob_s = []
for ckpt in ckpt_dir_s:
ckpt_dir = ckpt_dir_s_path + ckpt
if not os.path.isdir(ckpt_dir):
continue
nn = IcingIntensityNN()
nn.setup_eval_pipeline(data_dct, len(ll))
nn.build_model()
nn.build_training()
nn.build_evaluation()
nn.do_evaluate(ckpt_dir, ll, cc)
prob_s.append(nn.test_probs)
num = len(prob_s)
prob_avg = prob_s[0]
for k in range(num-1):
prob_avg += prob_s[k+1]
prob_avg /= num
probs = prob_avg
if NumClasses == 2:
preds = np.where(probs > prob_thresh, 1, 0)
else:
preds = np.argmax(probs, axis=1)
cc = np.array(cc)
ll = np.array(ll)
ice_mask = preds == 1
print(cc.shape, ll.shape, ice_mask.shape)
ice_cc = cc[ice_mask]
ice_ll = ll[ice_mask]
nav = GEOSNavigation(sub_lon=-75.0, CFAC=5.6E-05, COFF=-0.101332, LFAC=-5.6E-05, LOFF=0.128212, num_elems=2500,
num_lines=1500)
ice_lons = []
ice_lats = []
for k in range(ice_cc.shape[0]):
lon, lat = nav.lc_to_earth(ice_cc[k], ice_ll[k])
ice_lons.append(lon)
ice_lats.append(lat)
return filename, ice_lons, ice_lats
if __name__ == "__main__":
nn = IcingIntensityNN()
nn.run('matchup_filename')