From 9a9ddb53d4f7928d66a9b403da3bd9175f701bb2 Mon Sep 17 00:00:00 2001
From: tomrink <rink@ssec.wisc.edu>
Date: Thu, 20 Oct 2022 13:22:34 -0500
Subject: [PATCH] snapshot...
---
modules/deeplearning/icing_fcn.py | 12 ++++++------
1 file changed, 6 insertions(+), 6 deletions(-)
diff --git a/modules/deeplearning/icing_fcn.py b/modules/deeplearning/icing_fcn.py
index 07aa8adf..2cb96427 100644
--- a/modules/deeplearning/icing_fcn.py
+++ b/modules/deeplearning/icing_fcn.py
@@ -30,7 +30,7 @@ TRIPLET = False
CONV3D = False
NOISE_TRAINING = True
-NOISE_STDDEV = 0.10
+NOISE_STDDEV = 0.01
DO_AUGMENT = True
img_width = 16
@@ -566,7 +566,7 @@ class IcingIntensityFCN:
# activation = tf.nn.elu
activation = tf.nn.leaky_relu
- num_filters = len(self.train_params) * 10
+ num_filters = len(self.train_params) * 16
input_2d = self.inputs[0]
conv = tf.keras.layers.Conv2D(num_filters, kernel_size=5, strides=1, padding=padding, activation=None)(input_2d)
@@ -651,11 +651,11 @@ class IcingIntensityFCN:
conv = build_residual_block_1x1(conv, num_filters, activation, 'Residual_Block_2', padding=padding)
- conv = build_residual_block_1x1(conv, num_filters, activation, 'Residual_Block_3', padding=padding)
+ # conv = build_residual_block_1x1(conv, num_filters, activation, 'Residual_Block_3', padding=padding)
- conv = build_residual_block_1x1(conv, num_filters, activation, 'Residual_Block_4', padding=padding)
+ # conv = build_residual_block_1x1(conv, num_filters, activation, 'Residual_Block_4', padding=padding)
- conv = build_residual_block_1x1(conv, num_filters, activation, 'Residual_Block_5', padding=padding)
+ # conv = build_residual_block_1x1(conv, num_filters, activation, 'Residual_Block_5', padding=padding)
print(conv.shape)
@@ -681,7 +681,7 @@ class IcingIntensityFCN:
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 = int(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)
--
GitLab